Grimes, Samantha (2015) An evaluation of Aggression Replacement Training: the impact of a multi-component, CBT-based intervention on the problem behaviours, pro-social skills and moral development of pupils in English secondary schools. DAppEdPsy thesis, University of Nottingham.

Access from the University of Nottingham repository: http://eprints.nottingham.ac.uk/30479/1/samantha%20grimes%20doctoral%20thesis.pdf

Copyright and reuse:

The Nottingham ePrints service makes this work by researchers of the University of Nottingham available open access under the following conditions.

This article is made available under the University of Nottingham End User licence and may be reused according to the conditions of the licence. For more details see: http://eprints.nottingham.ac.uk/end_user_agreement.pdf

For more information, please contact eprints@nottingham.ac.uk

An Evaluation of Aggression Replacement Training: The impact of

a multi-component, CBT-based intervention on the problem

behaviours, pro-social skills and moral development of pupils in

English secondary schools.

Samantha Grimes

Thesis submitted to the University of Nottingham for the degree of Doctor of

Applied Educational Psychology

May 2015

2

Abstract

Aggression Replacement Training (ART) is a 10-week, multi-component

intervention based in cognitive behaviour therapy (CBT), which aims to improve

social competence. It has been applied internationally as part of offender

rehabilitation (NOMS, 2010). However, more recent research has focused upon

its application in school-based settings.

The aim of the current research is to investigate the efficacy of ART when

implemented in the UK with an adolescent sample in mainstream school

settings. These sessions were facilitated by newly-trained staff from the

Educational Psychology Service (EPS).

A quasi-experimental design was employed to evaluate this initial pilot of the

programme in one Local Authority. 41 participants across six settings were

allocated to intervention (N=23) and wait-list control (N=18) conditions. The

Social Skills Improvement System-Rating Scales (SSIS-RS), a multi-source

measure, was used to assess the group member’s problem behaviours and

social skills, with data gathered from teachers, parents and pupils themselves.

The Sociomoral Reflection Measure-Short Form (SRM-SF) was also completed

by the participants to ascertain their moral reasoning maturity.

Non-parametric statistical tests demonstrated no significant improvements in

the intervention participant’s social skills or problem behaviours. However, their

moral reasoning ability did increase significantly from pre to post-test, achieving

a large effect size (r=-0.64), which was not reflected in the data from the control

group.

In contrast to the quantitative findings, supplementary qualitative data gathered

from the facilitators and group members involved in the ART programme

demonstrated that all felt the intervention had resulted in positive outcomes for

the young people. Factors which may have contributed to the success of the

programme were also provided, including organisational support and group

composition.

3

Possible explanations for these findings, including methodological

considerations and comparison with previous research are discussed and the

implications of these findings in future practice and in guiding further research

suggested.

4

Acknowledgments

I would like to express my utmost gratitude to the following people for their

unerring support during the completion of this research project and my wider

doctoral training.

The tutor team at Nottingham University, particularly my supervisor Neil

Ryrie, for his guidance and patience throughout this challenging and

enlightening endeavour.

My fellow trainees, ‘TEP12’, for memories which will last a lifetime.

My colleagues from the EPS at which I have been lucky to spend my

final placement. With a special mention for Charlotte Reeve, Sandra

Lipkin and Deborah Benjamin my supervisors and Amy Ostrowski and

Gabrielle Pelter, my ‘inter-raters in chief’.

Professor Knut Gundersen for his guidance and insight into the world of

ART.

The staff at the schools and ART facilitators who contributed to this

research, for their participation, contributions and good humour.

And finally to my family and friends. My personal proof readers,

statisticians and emotional support. Especially Michael Glover, my ‘rock’

and Karen Grimes, to whom I owe a great deal.

5

Contents

Abstract ............................................................................................................. 2

Acknowledgments.............................................................................................. 4

1. Introduction .................................................................................................. 17

1.1 The Aim of the Research ........................................................................ 17

1.2 The Contribution to the Local Authority................................................... 18

1.3 The Unique Contribution of the Research ............................................... 18

1.4 The Researcher’s Personal Interest in this Area ..................................... 18

1.5 Overview of this Paper ........................................................................... 19

2. Literature Review ......................................................................................... 20

2.1 Introduction to the Literature Review ...................................................... 20

2.2 The National Context .............................................................................. 20

2.2.1 Government guidance ...................................................................... 20

2.2.2 Challenging Behaviour in Schools.................................................... 22

2.2.3 Exclusions ....................................................................................... 24

2.3 The Significance of Adolescence in the Development of Anti-Social

Behaviour ..................................................................................................... 25

2.3.1 Adolescence and Aggressive Behaviour .......................................... 25

2.3.2 Support to Develop Social Competence During Adolescence .......... 26

2.4 Aggression Replacement Training .......................................................... 26

2.4.1 Development of the Intervention ...................................................... 26

2.4.2 The Theory and Research Underpinning the Intervention ................ 27

2.4.3 The Efficacy of ART ......................................................................... 42

6

2.5 Implementation Science ......................................................................... 44

2.5.1 Implementation Science and ART .................................................... 46

2.6 A Systematic Literature Review of Aggression Replacement Training .... 47

2.6.1 Purpose of the Systematic Literature Review ................................... 47

2.6.2 Research Question for the Systematic Review ................................ 48

2.6.3 Methods Employed in the Systematic Review .................................. 48

2.6.4 Comparison of the Studies ............................................................... 50

2.6.5 In-depth Description of the Studies Found ....................................... 51

2.6.6 Quality Assessment ......................................................................... 55

2.6.7 Final Summary of the Findings ........................................................ 55

2.7 Introduction to the Following Evaluation of Aggression Replacement

Training ........................................................................................................... 58

2.7.1 Rationale for the Research .................................................................. 58

2.7.2 Research Questions and Hypotheses ................................................. 59

3. Methodology ................................................................................................ 61

3.1 Real World Research ............................................................................. 61

3.1.1 Evidence-Based Practice ................................................................. 62

3.2 Philosophical Stances: Ontology, Epistemology and Associated

Methodology ................................................................................................ 62

3.2.1. The Positivist Paradigm .................................................................. 63

3.2.2. The Constructivist Paradigm ........................................................... 63

3.2.3 Pragmatism ..................................................................................... 64

3.2.4 The Post-Positivist Paradigm ........................................................... 64

3.2.5 Epistemological Stance of the Current Research ............................. 67

7

3.3 The Research Project ............................................................................. 68

3.3.1 Design ............................................................................................. 68

3.3.2 Variables .......................................................................................... 70

3.4 Context and Participants ........................................................................ 70

3.4.1 Stakeholders .................................................................................... 70

3.4.2 Selection of Schools ........................................................................ 71

3.4.3 School Characteristics ..................................................................... 72

3.4.4 Sampling .......................................................................................... 73

3.4.5 Characteristics of the Participants .................................................... 75

3.5 Procedure .............................................................................................. 76

3.5.1 Setting up the Project in Schools ..................................................... 76

3.5.2 The Intervention Sessions ............................................................... 76

3.6 Measures Used ...................................................................................... 81

3.6.1 Measure of Problem Behaviours and Social Skills: The Social Skills

Improvement System ................................................................................ 81

3.6.2 Measure of Moral Reasoning: The Socio-Moral Reflection Measure-

Short Form ............................................................................................... 83

3.6.3 Data Collection Procedures ............................................................. 85

3.6.4 Additional Data ................................................................................ 86

3.6.5 Data Analysis ................................................................................... 87

3.7 Data Quality Issues ................................................................................ 87

3.7.1 Internal Validity ................................................................................ 88

3.7.2 External Validity ............................................................................... 92

3.7.3 Reliability ......................................................................................... 92

8

3.8 Ethical Considerations ............................................................................ 93

4. Results ........................................................................................................ 96

4.1 Quantitative Data Analyses .................................................................... 96

4.1.1 Dependent Variables and Direction of Change ................................ 96

4.1.2 Attrition ............................................................................................ 98

4.1.3 Descriptive Data .............................................................................. 99

4.1.4 Inferential Statistics ........................................................................ 108

4.1.5 Summary of Findings from Statistical Analyses.............................. 118

4.2 Qualitative Data Analysis...................................................................... 119

4.2.1 Qualitative Data Collection Procedures ......................................... 119

4.2.2 Thematic Analysis Process ............................................................ 121

4.2.3 Themes Constructed from the Group Member Interview Data ....... 123

4.2.4 Themes Constructed from the Facilitator Questionnaire Data ........ 130

4.2.5 Summary of Findings from Thematic Analysis ............................... 138

5. Discussion ................................................................................................. 140

5.1 Introduction .......................................................................................... 140

5.2 Reflections on Quantitative Findings .................................................... 140

5.2.1 Research Question 1 ..................................................................... 140

5.2.2. Research Question 2 .................................................................... 145

5.2.3 Consideration of the Quantitative Findings in Relation to Previous

Research ................................................................................................ 145

5.3 Reflections on Supplementary Qualitative Findings .............................. 151

5.3.1 Summary of Themes Constructed.................................................. 152

9

5.3.2 Linking the Themes Constructed to Theory and Previous Research

....................................................................................................... ........153

5.3.3 Summary ....................................................................................... 159

5.4 Methodological Limitations ................................................................... 160

5.4.1 Issues of Internal Validity ............................................................... 160

5.4.2 Issues of External Validity .............................................................. 163

5.4.3 Issues of Reliability ........................................................................ 163

5.4.4 Reflections on the Challenges Encountered in Real World Research

..................................................................................................................163

5.5 Implications of the Findings .................................................................. 165

5.5.1 Implications for Future Research ................................................... 165

5.5.2 Implications for Practice ................................................................. 167

5.6 Conclusions.......................................................................................... 169

5.6.1 Unique Contribution of the Current Research ................................ 169

5.6.2 Summary of Findings ..................................................................... 170

6. References ................................................................................................ 172

6.1 Secondary sources............................................................................... 193

7. Appendices ................................................................................................ 194

7.1 Appendix I: Systematic Review: A Detailed Description of the Process

Undertaken ................................................................................................ 194

7.2: Appendix II- Descriptive map of the studies to aid the in-depth review of

the research ............................................................................................... 196

7.3 Appendix III: Quality Assessment: A description of the Weight of Evidence

criteria employed to review the studies gathered from the systematic literature

search. ....................................................................................................... 200

10

7.4: Appendix IV Recruitment Leaflet and Initial Application Form from the

Educational Psychology Service to Support the Selection of Schools onto the

Project. ....................................................................................................... 203

7.5 Appendix V: Information Leaflet for Schools Regarding the Evaluation

Research Project. ....................................................................................... 206

7.6 Appendix VI: Information Letter and Consent Form for Parents ............ 212

7.7 Appendix VII: Information Letter and Consent Form for Teaching Staff

................................................................................................. ..................216

7.8 Appendix VIII: Information Letter and Consent Form for Participants in the

Experimental Group ................................................................................... 219

7.9 Appendix IX: Information Letter and Consent Form for Participants in the

Control Group ............................................................................................ 222

7.10 Appendix X: Information Letter and Consent Form for Educational

Psychologists Facilitating the Groups in the Experimental Condition. ......... 225

7.11 Appendix XI: Summary of the Structure of the ART

Sessions.......................................................................................................228

7.12 Appendix XII: Integrity Checklists from the international Center for

Aggression Replacement Training (iCART) ................................................ 230

7.13 Appendix XIII: The Socio-Moral Reflection Measure (Gibbs, Basinger &

Fuller, 1992) ............................................................................................... 237

7.14 Appendix XIV: The Prompt Sheet Used in the Semi-Structured

Interviews with the Participants in the Experimental Groups. ..................... 241

7.15 Appendix XV: Questionnaire Distributed to the Educational

Psychologists Facilitating the Experimental Groups. .................................. 242

7.16 Appendix XVI: Letter Confirming Ethical Approval Received from the

Ethical Committee of the University of Nottingham. .................................... 243

11

7.17 Appendix XVII Parental Information and Consent Form for the

Qualitative Measures ................................................................................. 244

7.18 Appendix XVIII: Pupil Information Sheet and Consent Form for

Participation in the Qualitative Measures ................................................... 246

7.19 Appendix XIX: Tables displaying exploratory analyses of normal

distribution .................................................................................................. 248

7.20 Appendix XX: Tables displaying the Results of the Analyses of

Homogeneity of Variances. ........................................................................ 254

7.21 Appendix XXI: Image of the Original Thematic Network for the Facilitator

Questionnaire Data .................................................................................... 256

7.22 Appendix XXII: Image of the Original Thematic Network for the Group

Member Interview Data .............................................................................. 258

7.23 Appendix XXIII: Needs of the Role Model and Target Individuals

Provided by the Member of Staff from Each School Responsible for Selection.

............................................................................................................. ......259

7.24 Appendix XXIV: Glossary of Abbreviations..........................................260

12

Tables

Table 2.1: Inclusion and Exclusion Criteria Employed During the Systematic

Review...............................................................................................................49

Table 2.2: Summary of the Results of the Quality Assessment of Studies

Identified by the Systematic Literature Review .................................................55

Table 3.1: Summary of Four Research Paradigms: Comparison of Ontology,

Epistemology and Methodology. Adapted from Fien, (2002); Guba & Lincoln

(1994. p.109) and Mertens, (2010)....................................................................66

Table 3.2: Key Characteristics of the Schools Involved in the Research. Data

gathered directly from the school’s databases as well as Ofsted’s (2013) ‘Data

Dashboard’.........................................................................................................72

Table 3.3: Participant Characteristics by Setting...............................................75

Table 3.4: A Week by Week Session Plan for Anger Control Training Taken

from Gundersen et al.

(2014.pp.84).......................................................................................................77

Table 3.5: Week by week session plan for the prosocial skills sessions (taken from Gundersen et al 2014. Pp. 31)...................................................................78

Table 3.6: Results of the Aggression Replacement Training Treatment Integrity Checks. .............................................................................................................80

Table 3.7: Table to Display the Subscales Included in the SSIS-RS.................82

Table 3.8: Results of SRM-SF Inter-Rater Analyses.........................................85

Table 3.9: Common Threats to Internal Validity and Actions Taken to Minimise

Their Effects in the Current Research. Adapted from Cohen, Manion & Morrison

(2007) and Babbie (2010)..................................................................................89

Table 4.1: Table to show the Dependent Variables Employed in the Study and

the Desired Direction of Change........................................................................96

Table 4.2: Table to Show the Number of Participants Present at Pre-test and

Post-test............................................................................................... ..............98

13

Table 4.3: Table to Show the Descriptive Data for the Self-report Measures at

Pre-test......................................................................................................... ...101

Table 4.4: Table to Show the Descriptive Data for the Self-Report Measures at

Post-test..................................................................................................... ......102

Table 4.5: Table to Show the Descriptive Data for the Teacher Report Measures

at Pre-test........................................................................................................103

Table 4.6: Table to Show the Descriptive Data for the Teacher Report Measures

at Post-test.......................................................................................................104

Table 4.7: Table to Show the Descriptive Data for the Parent Report Measures

at Pre-test........................................................................................................105

Table 4.8: Table to Show the Descriptive Data for the Parent Report Measures

at Post-test.................................................................................................. .....106

Table 4.9: Table to Show the Significant Results of Between-group Pre-test

Analyses..........................................................................................................108

Table 4.10: Table to Show the Significant Results of Between-group Post-test

Analyses..........................................................................................................110

Table 4.11: Table to Show the Significant Changes from Pre to Post-Test from

the Within-Group Analyses..............................................................................114

Table 7.1: Table to Show the Results of the Literature Search Conducted on 4th

and 5th June 2014............................................................................................195

Table 7.2: Descriptive Map of the Studies Found in the Systematic Literature

Review....................................................................................................... ......196

Table 7.3 Table to Show the Skew, Kurtosis and Shapiro Wilk Analyses for the

Self Report Data from the Control Group.........................................................248

Table 7.4: Table to Show the Skew, Kurtosis and Shapiro Wilk Analyses for the

Self Report Data from the Intervention Group.................................................249

14

Table 7.5: Table to Show the Skew, Kurtosis and Shapiro Wilk Analyses for the

Teacher Report Data from the Control Group..................................................250

Table 7.6: Table to Show the Skew, Kurtosis and Shaprio Wilk Analyses for the

Teacher Report Data from the Intervention Group..........................................251

Table 7.7: Table to Show the Skew, Kurtosis and Shaprio Wilk Analyses for the

Parent Report Data from the Control Group....................................................252

Table 7.8: Table to Show the Skew, Kutosis and Shapiro-Wilk Analyses for the

Parent Report Data for the Intervention Group ...............................................253

Table 7.9: Table to Show the Non-parametric Levene’s test Results for the Self

Report Measures.............................................................................................254

Table 7.10:Table to Show the Non-parametric Levene’s test Results for the

Parent Report Measures..................................................................................255

Table 7.11: Table to Show the Non-parametric Levene’s Test Results for the

Teacher Report Measures...............................................................................255

Table 7.12 Behaviour Descriptors for the Role Models and Target Pupils

Provided by School Staff During the Selection Process..................................259

Table 7.13 Table of Abbreviations...................................................................260

15

Figures

Figure.2.1. A Multi-Level Model of Factors that Influence Implementation

Quality. (Domitrovich et al. 2008. p.8)................................................................46

Figure.4.1: Graph to show Changes in Median SRMS from Pre to Post-Test for

the Control and Intervention Groups ...............................................................112

Figure 4.2: Graph to show Changes in Median Self-Reported Internalising

Score from Pre to Post-Test for the Control and Intervention Groups ............112

Figure 4.3: Graph to Show Changes in Median Teacher-Reported Externalising

Score from Pre to Post-test for the Intervention and Control Groups..............113

Figure 4.4: Graph to Show the Changes in Median Parent Reported Autism

Spectrum Scores from Pre to Post-Test for the Intervention and Control

Groups.............................................................................................................113

Figure 4.5: Graph to Show the Positive Relationship Between Teacher-

Reported Cooperation Change Scores and Percentage Attendance at

Sessions..................................... ....................................................................117

Figure 4.6: Graph to Show the Positive Relationship Between Parent-Reported

Autism Spectrum Change Scores and Percentage Attendance at

Sessions..........................................................................................................118

Figure 4.7: Thematic map of the overarching theme ‘Reported Outcomes

Experienced’ and the associated themes generated from group member

interview data in response to the question: What are the views and experiences

of those involved in the initial pilot of the ART intervention sessions, in relation

to programme implementation, contents and

effectiveness?..................................................................................................124

Figure 4.8: Thematic map of the overarching theme ‘Perceptions of Intervention

Contents and Implementation’ and the associated themes generated from group

member interview data in response to the question: What are the views and

experiences of those involved in the initial pilot of the ART intervention

sessions, in relation to programme implementation, contents and

effectiveness?..................................................................................................125

16

Figure 4.9: Thematic map of one overarching theme and the associated themes

generated from facilitator questionnaire data in response to the question: What

are the views and experiences of those involved in the initial pilot of the

Aggression Replacement Training intervention sessions, in relation to

programme implementation, contents and effectiveness?...............................130

Figure 4.10: Thematic Map of the Second Overarching Theme and Associated

Themes Generated from Facilitator Questionnaire Data in Response to the

Question: What are the views and experiences of those involved in the initial

pilot of the Aggression Replacement Training intervention sessions, in relation

to programme implementation, contents and

effectiveness?..................................................................................................131

Figure 7.1: A Flow Chart Depicting the Search Strategy Employed in the

Systematic Review...........................................................................................194

Figure 7.2 Original Thematic Network for the ‘Reported Outcomes’ Overarching

Theme for the Facilitator Questionnaire Data..................................................256

Figure 7.3 Original Thematic Network for the ‘Factors Impacting Upon Success’

Overarching Theme for the Facilitator Questionnaire Data.............................257

Figure 7.4: Original Thematic Network for Both Overarching Themes

Constructed from the Group Member Interview Data......................................258

17

1. Introduction

1.1 The Aim of the Research

The aim of this study is to evaluate an ART intervention, when implemented in

English secondary schools.

ART (Glick & Gibbs, 2011; Goldstein et al 1987 & Glick; Goldstein, Glick &

Gibbs, 1998) is a ten week, multi-component intervention based on the

principles of CBT. The programme aims to improve skills of social competence

and reduce aggression, which is viewed by the authors as ‘...an overt behaviour

often employed by those weak or lacking in pro-social alternatives’ (Goldstein,

Glick & Gibbs, 1998, p.1). The intervention adopts a multimodal approach, as

aggression is viewed as consisting of multiple interlocking internal and external

causes, ‘...a behavioural, cognitive and emotional phenomenon...’ (Glick &

Gibbs, 2011, p.3).

Whilst the current programme is primarily aimed at adolescents (Glick & Gibbs,

2011), ART has been adapted for a range of age groups and special

populations (Gundersen et al. 2014). Evaluations have involved adult offenders

(Hatcher et al. 2008; Sugg, 2000 cited by McGuire & Clark, 2004); adults with

learning disabilities (Curulla, 1991) and young people with Autism (Moynahan,

2003). The programme has also been applied in a residential centre for the

treatment of behavioural disorders (Coleman, Pfeiffer & Oakland, 1992); youth

justice custodial settings (Currie et al. 2012); runaway shelters (Nugent, Bruley

& Allen, 1998/1999) and in secure centres for young offenders (Erickson, 2013;

Glick & Goldstein, 1987; Holmqvist, Hill & Lang, 2009; Roberts, 2009). More

recent studies across Europe, North America and Australia have begun to

investigate the impact of ART when implemented in school settings (Gundersen

& Svartdal, 2006; Gundersen & Svartdal, 2010; Jones, 1991; Langeveld,

Gundersen & Svartdal, 2012; Moynahan & Stromgren, 2005; Novy &

McFarland, 2011).

The, predominantly positive, findings mean that ART was awarded model

programme status from the Office of Juvenile Justice and Delinquency

Prevention (US Department of Justice, n.d.). In 2001 ART also became an

18

accredited programme for use in the probation and prison services in England

and Wales (McGuire & Clark, 2004) and is now well-established within the UK

Criminal Justice System (National Offender Management Service, NOMS,

2010).

1.2 The Contribution to the Local Authority

The EPS with which I am placed invested in the ART programme as part of a

new initiative which contributed towards the service’s development plan. This

aimed to reduce anti-social behaviour in schools, as well as prevent possible

future criminal behaviour. As the intervention package required substantial

commitment and resources from the EPS, it was decided that it would be

beneficial to obtain an indication of the programme’s effectiveness before it was

made available to schools more widely. Therefore the following research

focuses upon the initial, small-scale implementation and evaluation of the

programme, which provided the Educational Psychologists (EPs) with an

opportunity to practise as newly qualified facilitators.

1.3 The Unique Contribution of the Research

The studies mentioned above provide a strong evidence base regarding the

outcomes of ART internationally. A recent survey conducted by the London

Probation Trust found that ART was implemented in at least 10 countries

worldwide (NOMS, 2010). However, despite evidence suggesting that ART has

been utilised within the probation service for over a decade (McGuire & Clark,

2004), there is a distinct lack of research into ART in England. At the time of

writing, there is also no existing evidence regarding the effectiveness of ART,

when implemented to support young people with social competence needs, in

the UK. The transferability of the ART programme to children and young people,

given its foundation in adult work, will be discussed later in Section 2.

1.4 The Researcher’s Personal Interest in this Area

The area of Special Educational Need (SEN) referred to as Behavioural,

Emotional and Social Difficulties (BESD) has always been a great interest of

mine, especially in regard to improving outcomes for young people with such

needs. This began with my previous occupations as a teacher and as an

inclusion support worker for a Local Authority BESD support team, conducting

19

therapeutic work with young people and advising school staff. My interest and

knowledge in this area has since been nurtured by my current role as a Trainee

EP. Working in education, I was disheartened by the overwhelming tendency for

behaviour difficulties to be viewed negatively by those responsible for these

young people, at a time when learning needs were being viewed from a more

positive, inclusive standpoint.

The importance of evidence-based practice and the role of the EP in

contributing to such a knowledge base had been instilled during my doctoral

training. When embarking upon the research I explored several possible

evaluation projects, as I was inspired to contribute to the evidence-base

surrounding interventions implemented in education settings, but it was my

interest in the above areas which drove me to contact the EP who was spear-

heading the current project, as she had previous research experience and

personal interest in the areas of anti-social behaviour and gang culture.

1.5 Overview of this Paper

Chapter 2: The Literature Review provides a brief summary of the national

context in which the research is conducted. A description of the Aggression

Replacement Training programme is then provided along with an overview of

associated theory and research.

Chapter 3: The Methodology section provides a detailed account of the design

and implementation of the current research. Consideration is also given to

ethical and methodological issues and the possible impact that these may have

on the findings generated.

Chapter 4: The Results section presents the findings from the statistical

analyses in order to judge the significance of the results. This is followed by

thematic analyses of the supplementary qualitative data gathered.

Chapter 5: Finally during the Discussion the research questions will be

reviewed in light of the findings of the current research. Limitations of the study

will be highlighted and the consequences discussed. A conclusion will then

provide an overall summary of this research project.

20

2. Literature Review

2.1 Introduction to the Literature Review

The aim of the following literature review is to provide the reader with an

understanding of the theory and research evidence associated with ART.

The literature review begins by describing the context in which this research is

being undertaken, which highlights the importance of early, targeted,

intervention for adolescents displaying behaviour considered to be aggressive.

An overview of the theory behind the ART intervention and the constructs it

aims to change is provided. This is followed by a brief summary of

Implementation Psychology, an area pertinent to evidence-based practice and

programme evaluation. Finally, a systematic review of previous research

evaluations into ART is conducted in order to examine the existing evidence

base and inform the current research.

2.2 The National Context

2.2.1 Government guidance

2.2.1.1 Focus on Improving Mental Health and Wellbeing.

Mental Health refers to a child’s ability to develop intellectually, spiritually and

emotionally, so that they can make the most of the opportunities and

relationships that they encounter (Mental Health Foundation, 1999). Mental

health difficulties can manifest in many ways including behaviours perceived as

aggressive and antisocial (Mental Health Foundation, 1999). According to a

recent survey conducted by UNICEF, compared to 20 other economically

developed countries, the United Kingdom scored within the bottom third on five

of the six dimensions of child wellbeing measured (UNICEF, 2007). The three

dimensions in which the country came last were family and peer relationships,

behaviours and risks, and subjective wellbeing. This suggests that relating to

others and avoiding risky or violent behaviour are areas which require a great

deal of support if the wellbeing of British children is to improve.

21

The Every Child Matters Framework (DfES, 2003) highlighted the importance of

promoting pupil’s emotional health and wellbeing in order to improve their life

chances and reduce the number who ‘...engage in offending or anti-social

behaviour...’ (DfES, 2003, p.5). Five outcomes for all children and young people

were identified as aims for the strategy:

Being healthy;

Staying safe;

Enjoying and achieving;

Making a positive contribution;

Achieving economic wellbeing.

This initial Green Paper provided a starting point for further publications

providing guidance on supporting the mental health and wellbeing of young

people (HMSO, 2007) and improving behaviour in schools (DfES, 2005; Ofsted,

2005). These principles are also reflected in the new Special Educational Needs

Code of Practice which states that education should enable all children to live

‘...fulfilling lives...’ (DfE, 2015, p92).

2.2.1.2 The Inclusion Agenda

In an attempt to secure consistency and equity regarding the education of

children with SEN, The Warnock Report (DES, 1978) advocated integration. All

children should be educated together so as to experience the same

opportunities to succeed. This principle was secured in legislation by the

Education Act 1981. Government guidance has also stressed that inclusion

means that children are educated alongside each other, not segregated in

special units but participating fully in the life of the school (DfEE, 1997). This

notion of ‘full participation by all’ (BPS, 2005, p.1) has continued to be a focus of

government publications ever since.

Several threats to the realisation of being fully inclusive still remain. Firstly

current legislation (Special Educational Needs and Disabilities Act, SENDA,

2001), maintains a caveat which suggests that students have to be educated in

mainstream settings, as long as this does not disrupt the learning of others. This

may imply that externalised behavioural difficulties require additional support,

beyond that which mainstream school can provide. There is also evidence that

22

teacher’s unions hold concerns that inclusion increases the type of students

who lower teacher credibility (Tomlinson, 2005). Accountability for educational

results may also lead to the exclusion of pupils who are perceived to threaten

the school’s image of functioning (Stirling, 1991).

2.2.2 Challenging Behaviour in Schools

Whilst the behaviour in the majority of schools in England is reported to be

‘good’ or even ‘outstanding’, there remains to be children whose behaviour

disrupts progress and concerns school staff (Ofsted, 2005). Classroom

misbehaviour includes a wide range of conduct which disrupts learning, from

low level disturbance to behaviour defined as physical assault and aggression

(DfE, 2012). Teachers often consider difficult behaviour to be one of the most

challenging aspects of their work (Merrett and Whedall, 1987) possibly because

it can be interpreted as a threat to their authority (Gray, Miller & Noakes, 1994).

According to a recent survey conducted by a teacher’s union in the United

Kingdom, 86% of teachers had to deal with challenging behaviour during the

academic year and a third experienced physical violence from pupils in school

(ATL, 2012).

2.2.2.1 Behavioural, Emotional and Social Difficulties

Defining BESD is problematic (Frederickson & Cline, 2009). It would appear

that the children in this category do not form a homogenous group and

represent a wide range of difficulties (Kershaw & Sonuga-Barke, 1998). A

continuum approach is proffered, from behaviour which is disruptive but falls

within normal bounds to behaviour which is thought to be indicative of a serious

mental health issue (DfE, 1994).

The term ‘Behavioural’ in this category has recently been replaced (DfE, 2015,

p.85), to become Social, Emotional and Mental Health. Here it is suggested that

poor emotional and social development may lead to challenging and disruptive

behaviour because of immature social skills. These skills deficits may manifest

as conduct problems such as aggression or cause the child to withdraw socially

and develop mood disorders.

According to figures from the Department for Education (DfE, 2013a)

Behavioural, Emotional and Social Difficulty is the most common category of

23

SEN in secondary schools. Whilst the majority of this group of young people

continue to make the expected progress in primary school, this figure decreases

considerably between key stage 2 and 4 (DfE, 2013a). For over a decade the

government have recognised that this group is a priority, in terms of securing

early appropriate support, to avoid some of these negative outcomes (DfEE,

1997). This report clearly stated that responsibility for meeting the needs of

such pupils rests with mainstream schools.

2.2.2.1.1 The Social Construction and Systemic Basis of Aggression

Whilst a thorough critical review is beyond the scope of this thesis, it is

important at this point to consider the constructed nature of behaviour, or more

specifically aggression, which is the social concept focused upon in this report.

Whether or not behaviour is labelled ‘aggressive’ depends on the judgement of

the observers of that behaviour. The criteria applied in this decision making

process differ widely, including the characteristics and intensity of the behaviour

and the intentions attributed to the performer by the observer, making it a

‘...socially defined...’ concept (Bandura, 1973, p.8).

Whilst making judgements about the behaviour of others serves an evolutionary

purpose, allowing us to predict a person’s future behaviour by labelling their

actions, these beliefs can also be wrongly attributed and harmful, impacting

upon other’s expectations of that young person and their own self-concept. The

social interactionist perspective would suggest that aggression is the result of

situational and interpersonal factors (Felson & Tedeschi, 1995), whilst social

learning approaches describe aggression as arising from observation and

modelling of other’s behaviour (Bandura, 1973). These explanations make it

unfeasible to describe an individual as ‘aggressive’ as they assert that context

plays a considerable part in the resultant behaviour. Therefore to avoid referring

to the behaviour of young people in a deterministic manner, it will henceforth be

referred to as an observation or perception of others.

In keeping with this theme, theorists moving away from a within-child model of

needs advocate a focus upon the context and interpersonal relationships which

instigate and perpetuate these difficulties (Cooper, Smith & Upton, 1995; Maras

& Kutnick, 1999; Miller, 2003). In practice, EPs are encouraged to adopt a more

24

holistic approach during casework to incorporate social context and other

situational factors into their assessments and the interventions devised (Cline,

1992; Frederickson & Cline, 2009), using models such as Bronfenbrenner’s

(1996) Ecosystemic model to guide them. Ecological approaches such as this

view aggression as a learned response due to patterns of interaction between

the young person and the environments which they are embedded in

(Frederickson & Cline, 2009). Such an assessment would provide details to

inform a more holistic and arguably more ethical intervention package than

traditional ‘child-deficit’ support, with a focus on influencing the context as well

as supporting the individual in developing their skill-base (Kelly, 2008).

2.2.3 Exclusions

2.2.3.1 Who is at Risk?

According to recent Government statistics (DfE, 2013b) the number of

permanent exclusions in 2011/2012 rose, following a steady decrease since

2002/2003. However, the number of fixed term exclusions continued on its

downward trend. The most cited reason for exclusion in both instances was

persistent disruptive behaviour and the most vulnerable age is between 13 and

14 years old. Young people with Behavioural, Emotional and Social needs were

the most likely to be excluded. Despite guidance stating that ‘...every

practicable means to maintain the pupil in school...’ should be exhausted

(DCSF, 2008, p.28).

2.2.3.2 What are the Causes of Exclusion?

Attwood, Croll and Hamilton (2003) interviewed 26 young people who had been

excluded from mainstream education. Difficulties with personal relationships

were the most commonly cited reason for the exclusion. Similarly, a large scale

survey conducted by Costenbader and Markson (1998) found that antisocial

behaviour, specifically physical aggression, was the most prevalent cause cited

by students for their suspension. The participants stated that they would have

liked the opportunity to learn alternatives to the negative behaviours for which

they received the exclusion and only 19% believed that the suspension had

helped them and had led to changes in their behaviour. This suggests that

25

exclusion is not an effective method of behaviour management, which is a view

supported by others in this area (McGinnis, 2003).

2.2.3.3 The Negative Outcomes Associated with Exclusion

Prolonged periods of time away from school may not only restrict access to the

curriculum and disrupt the child’s education but also removes the opportunity for

social interaction (Gersch & Nolan, 1994). The Office of the Children’s

Commissioner (2012) concluded that schools should only resort to exclusion as

a last resort, after alternative methods have been attempted. Research has

found that even shorter periods of exclusion, or ‘suspension’, are associated

with lowered academic gains and withdrawal from education completely (Arcia,

2006) and that often attempts at re-integration into mainstream are not

successful (Daniels et al., 2003). Exclusions have been associated with

patterns of offending and substance misuse (Stirling, 1991) and are also costly

to society. Parsons and Castle (1998) found that providing additional resources

for those considered to display behavioural difficulties was cheaper than

exclusion both financially and in regard to the impact on the individuals

involved.

2.3 The Significance of Adolescence in the Development of Anti-Social

Behaviour

2.3.1 Adolescence and Aggressive Behaviour

The previous section highlighted the significance of adolescence, in relation to

the possible development of disruptive, anti-social behaviour which can lead to

negative outcomes should the child experience exclusion from education.

Antisocial behaviour in the form of perceived deviant and violent behaviour is

found to peak during adolescence before beginning to decline after age 11

(Bjorkqvist, Lagerspetz & Kaukiainen, 1992). Further evidence highlighting the

significance of this stage of life, in relation to the development of challenging

behaviour, comes from research which suggests that first criminal offences

often occur between the age of 11 and 12 and that most youth offenders are

between the ages of 14-16 years (Philips & Chamberlain, 2006). One possible

explanation is that neurological changes which occur during adolescence have

26

been associated with the impulsive, risk-taking behaviours observed during this

period (DfE, 2011a).

2.3.2 Support to Develop Social Competence During Adolescence

Recent government policy (DfE, 2011b) highlights the importance of ensuring

adequate support for young people during this stage of life in order to ensure

they fulfil their potential ‘...through positive relationships....’ (p.ix). It is important

for young people to learn patterns of behaviour that enable them to contribute to

society in a positive manner (Csikszentmihalya & Larson, 1984) including

social-emotional skills (DCSF, 2007). Guidance appears to suggest that

responsibility for providing early, appropriate support and exhausting every

avenue to ensure the pupil continues with their education lies with the schools

(DCSF, 2008; DfEE, 1997)

The following section describes a multi-component intervention programme,

devised to be implemented within school settings, with the intention of

supporting the development of social competence in this ‘at risk’ population.

2.4 Aggression Replacement Training

2.4.1 Development of the Intervention

ART was originally developed in the 1980s in response to the high rates of

youth crime in America (Goldstein et al 1987). The initial pilot studies,

conducted in the USA, led to the conclusion that ART is an effective psycho-

educational intervention for young people considered to show aggression in

youth custodial settings (Glick & Goldstein, 1987).

Over the years ART has gone through several transformations (Goldstein, Glick

& Gibbs, 1998; Glick and Gibbs, 2011). One of the current versions was created

in Norway for use in European schools (Gundersen et al. 2014) by researchers

from the international Centre for Aggression Replacement Training (iCART), the

European training provider of ART. It is based on the observations of staff at

child welfare organizations and their extensive experience in implementing ART

but the content remains the same as that provided by Goldstein, Glick and

Gibbs (1998). However, the delivery of some sessions has been modified (see

27

section 3.5.2 and Appendix XI for further detail regarding the contents and

structure of the intervention sessions).

2.4.2 The Theory and Research Underpinning the Intervention

This section begins by describing the three components of the ART

intervention. This is then followed by a discussion surrounding the theoretical

basis of the intervention, CBT, before a final exploration of the two concepts

which ART aims to modify; aggression and social competence.

2.4.2.1 The Three Components of the Intervention

ART is a multimodal programme. Whilst each component is considered

effective individually, it is the combination of all three elements which are

believed to promote long term success (Glick & Gibbs, 2011). The following

section will outline the role of each component in contributing to the reduction of

aggression and nurturing of social competence.

2.4.2.1.1 Pro-social Skills Training

This component aims to provide adolescents with pro-social skills to use in

place of behaviours perceived as aggressive.

According to the Sensorimotor Skill Model of social interaction (Argyle &

Kendon, 1967), socially skilled behaviour is comprised of 3 interrelated

components: Social perception, social cognition or ‘Translation Processes’

(p.58) and social performance. Being able to interpret other’s feelings,

intentions and actions is important as our perception will affect which response

we believe is appropriate and the actions we then carry out (Argyle, 1994).

In the skill deficit model advocated by Goldstein et al (1980), youngsters

described as aggressive are perceived as being weak in interpersonal skills and

enhanced levels of social cognition (Goldstein, Glick & Gibbs, 1998), which is

remediated through explicit training. This includes issues of perception and

social information processing such as misinterpretation of cues, biased

attribution of intent and deficient social problem solving (Lochman & Dodge,

1994; Zelli et al. 1999). Such young people also appear to suggest more

agonistic strategies when dealing with social dilemmas and are more likely to

actually enact such behaviours when attempting to reach their goals (Rubin,

28

Bream & Rose-Krasnor, 1991). It has been suggested that adopting a single

approach does not appear to result in long term behaviour change and

therefore programmes which address both behaviours and cognitions are

advocated (Frederickson and Simms, 1990; Pepler, King & Byrd, 1991).

Intervention evaluation studies have found that explicit skill training is not only

successful in increasing the positive social behaviours of young people

(Denham et al. 2006) but also in reducing the behaviours considered to be

antisocial, such as aggression (Pepler, King & Byrd, 1991). However, not all

findings have been so positive, in a meta-analysis of 35 studies; Quinn et al

(1999) found a mean effect size of 0.199 for social skills training groups for

children with BESD. They suggested that the interventions would have been

more effective if they had been tailored to the needs of the participants.

However, Gresham et al (2004) assert that the studies in Quinn’s review did not

possess high validity ratings, nor did they all use outcome measures of social

behaviour or studies where all participants had BESD. These authors describe

four alternative, stringent meta-analyses which found more positive results,

leading to the conclusion that social skills training is beneficial for pupils with, or

at risk of, BESD.

The social skills training component of ART is ‘...a systematic, psycho-

educational intervention to teach pro-social behaviours.’ (Glick & Gibbs, 2011.

p.14). A core curriculum of 10 different social skills is provided. However, in

keeping with a prescriptive model of intervention, these can be adapted based

on the needs of those in the group (Gundersen et al. 2014). The sessions use

strategies based in social learning theory such as modelling, role play and

feedback (Gundersen et al. 2014) to develop skills which serve to displace the

behaviour considered to be destructive (Hollin, 2004) as well as opportunities

for open reflection of thoughts, in keeping with CBT principles (Gundersen et al.

2014)

The success of intervention groups depends on many factors, Tierney and

Dowd (2000) suggest three conditions; ensuring the work is valued by the

school, giving facilitators time to develop and reflect on their skills and giving

pupils the choice to take part. The participant’s acceptance of the skills being

29

taught, in terms of their face validity, is also an important factor to consider

(Sarason & Sarason, 1981).

2.4.2.1.2 Anger Control Training (ACT)

This component supports young people in inhibiting anger, an emotion which

can interfere with the adoption of pro-social behaviour (Glick & Gibbs, 2011).

Anger, ‘...an emotional response to provocation...’ (Novaco, 1975, p.5), is

thought to consist of both physiological arousal and cognitive appraisal of those

feelings as anger (Indoe, 1995). Aggression may arise from concomitant

thoughts, such as unmet expectations or reduced inhibition and impulsive

behaviour arising from the increased emotional arousal (Novaco, 1975).

Goldstein, Glick and Gibbs (1998) suggest that a tendency to employ

aggressive means to achieve personal goals may be indicative of anger control

problems. Certainly, research has found that high levels of anger relate to

increased frequency of perceived aggressive outbursts in young people

(Cornell, Petersen & Richards, 1999).

Traditional behaviour modification strategies for individuals considered to show

aggression do not appear to lead to generalisation of behaviours into natural

environments. Instead theorists highlight the importance of considering

emotions and their role in explaining difficult behaviour (Faupel, 2002) and the

associations between cognitive distortions and overt behaviour, such as

aggression (Barriga et al. 2000; Beck, 1999; Liau, Barriga & Gibbs, 1998;

Lochman & Dodge, 1994). Therefore researchers began to focus on

remediating the emotional arousal which often underpinned aggression, anger,

by restructuring maladaptive thought processes alongside teaching alternative

behaviours (Feindler & Baker, 2004).

In order to improve self-control, which is related to aggression in later life (Caspi

et al. 1995), structured group anger management programmes utilise a range of

different techniques, including training in social skills and cognitive-relaxation

coping. These have been found to be equally successful (Deffenbacher et

al.1994). Such interventions have also achieved some success when aimed at

young people within educational settings (Dwivedi & Gupta, 2000; Feindler &

Baker, 2004; Lochman & Wells, 2003). However, it has been suggested, similar

30

to pro-social skill training, that tailoring of the interventions to the needs of the

individuals may prove most beneficial (Edmondson & Conger, 1996). For

example, if an individual does not have knowledge of pro-social skills, cognitive

restructuring and relaxation will not provide them with such skills.

The theory behind the techniques used in this component of the intervention

stems from the work of Luria (1961) who investigated children’s use of inner

speech in controlling their external behaviours. Meichenbaum and Goodman

(1969a) continued this work by finding evidence of a developmental sequence

of verbal control of behaviour. The most effective form of verbal control became

more covert as children developed. Whilst younger children required overt

verbalisations for optimal functioning, older children were hindered by being

forced to verbalise aloud and found covert instructions more helpful. Impulsive

children have been found to possess less verbal control over their behaviour,

using their speech less efficiently (Meichenbaum & Goodman 1969b). Cognitive

self-instruction training, which nurtures overt and covert strategies including

questioning, planning, self-guidance and self-evaluation, reduced impulsivity

and improved reflection, giving children control over their behaviour

(Meichenbaum & Goodman, 1971).

In keeping with the cognitive-behavioural self-control approach, Novaco (1975)

believed that an indirect link exists between the provoking event and anger,

which is mediated by cognitive appraisal of the event. Anger was believed to be

‘...fomented, maintained and inflamed by the self-statements that are made in

provocation situations.’ (p.23). Novaco’s treatment procedures involved both

cognitive control procedures and relaxation techniques in order to alleviate the

physiological response and improve self-control. Following the intervention,

participant’s management of anger significantly improved compared to controls.

It is important to note that some individuals employ aggression to reach

instrumental goals that are not fuelled by anger, these individuals would not

necessarily be suited to ACT (Novaco, Ramm & Black, 2004).

Novaco’s programme was developed into a sequence of taught techniques by

Eva Feindler (Feindler, 1995; Feindler & Ecton, 1986). As anger involves an

interaction of physiological, behavioural and cognitive factors (Faupel, Herrick &

31

Sharp, 2011), a combination of techniques are employed here to address all

three. Research suggests that this package lowered externalising behaviour in

adolescents when implemented in a home for youth with delinquent behaviour

(Nugent, Champlin & Wiinimaki, 1997) and in school settings (Feindler, Marriott

& Iwata, 1984; Whitfield, 1999). This intervention provides the basis for the ACT

component of ART. During ACT sessions participants are taught a chain of

strategies including recognition of perceptions and triggers, interpretation of

cues of arousal, reduction of arousal and the use of self-instructional reminders,

consideration of consequences and self-evaluation (Glick & Gibbs, 2011).

Throughout these stages cognitive restructuring techniques are used to identify

irrational thought patterns and replace them with a more rational appraisal of

the situation (Gundersen et al. 2014).

2.4.2.1.3 Moral Reasoning Development

The final component of ART aims to increase moral values, making the

individual more likely to employ the pro-social skills taught previously (Gibbs,

2004).

Morality develops over time, depending on the child’s social experiences and

cognitive ability to process these experiences (Guerra & Bradshaw, 2008). Both

Kohlberg (1973/1984) and Piaget (1965) linked socialization to moral

development. They claimed that moral reasoning ability develops sequentially,

through social experiences and ‘role-taking opportunities’ (Kohlberg, 1984,

p.199), particularly with peers. In Kohlberg’s theory this equates to three levels

of moral development, each with two stages. Children progress from making

moral judgements based on avoiding punishment and being concerned with

their own needs (stage 1 and 2), to adhering to social norms and laws (stage 3

and 4), until finally internalising principles such as justice, equality and dignity

(stage 5 and 6). Research confirms that children and adolescents progress

through these stages sequentially (Colby et al. 1983). It would also appear that

the development of moral reasoning, specifically the shift from stage 2 to stage

3 thinking in late childhood and early adolescence, is universal, being found

across cultures and measures (Gibbs et al. 2007).

32

Kohlberg (1973) applied his theory to antisocial behaviour, posing that those

perceived to act in a ‘delinquent’ manner possessed a delay in moral

development, so that the immature ‘pre-moral’ stages of 1 and 2 persist into

adolescence. This implies that individuals lack the appropriate processes to

control their behaviour.

Kohlberg’s theory was not without criticism, for example few reach post-

conventional stages and those who do tend to originate from western, urban

cultures (Snarey, 1985). Also young children have been found to possess a

deeper sense of morality than Kohlberg proposed (Damon, 1999).

Gibbs’ model of socio-moral development (Gibbs, Bassinger & Fuller, 1992),

adapted from Kohlberg, comprises of 4 stages. Moral judgement at stage 1 and

2 is immature and superficial, focusing on power, consequences and deals,

whereas stage 3 and 4 are mature and profound, concerned with mutuality,

interpersonal expectations and the good of society. These stages develop

across the lifespan, with stage 1 commonly associated with young children and

stage 4 late adolescence and adulthood (Gibbs, 2010). Therefore adolescents

who remain at stages 1 or 2 are considered delayed (Glick & Gibbs, 2011).

In support of these developmental theories researchers have concluded that the

moral reasoning of adolescents who have committed criminal acts is immature

in comparison to those who have not (Palmer & Hollin, 1998; Stams et al.

2006). Specifically the former group of adolescents have been found to mostly

employ Gibbs’ stage 2 reasoning, compared to the latter who use stage 3

(Gregg, Gibbs & Basinger, 1994). More mature moral reasoning has been found

to relate to greater pro-social behaviour such as helping (Eisenberg et al 1991).

The delay, often arising from a lack of opportunity to take other’s perspectives,

implies that not only are young people who engage in behaviours considered to

be antisocial at an immature moral stage, but they also possess persistent

egocentric bias or self-serving cognitive distortions, characteristic of much

younger children (Gibbs, 2004). These faulty beliefs correlate with the perceived

aggressive and antisocial behaviours engaged in, for example individuals

considered to be aggressive will justify an act, such as stealing, by saying it

fulfilled their needs and not consider other’s perspectives (Gibbs, 2010).

33

Research supports Gibbs’ theory, in that self-serving distortions have been

found to be elevated in adolescent offenders (McCrady et al. 2008) and

correlate positively with problem behaviours observed in college students

(Barriga et al. 2001).

A child does not possess moral values purely because they can affirm them.

They also need to be able to understand why they are important (Gibbs, 2004).

For example, both offenders and non-offenders state that moral values are

important but then differ in the maturity of their moral reasoning (Palmer &

Hollin, 1998). Kohlberg believed that mature moral judgements could be

stimulated through education in the form of classroom discussion programmes

(Kohlberg 1973).

Intervention evaluation research has found that moral dilemma discussion

groups, aimed at adolescents considered to show behaviour difficulties within

school contexts, have been found to be highly successful (Arbuthnot & Gordon,

1986). Not only did participant’s moral reasoning ability mature significantly

compared to control pupils, but a decrease in problem behaviours observed

was also reported following the sessions. The researchers concluded that

developing the basis for decisions improved behaviour. Although this change in

cognition does require some support from others if it is to be maintained and

have a sustained influence on behaviour (Gibbs et al. 1984).

In ART moral values are encouraged by providing the young people with

opportunities to adopt the perspectives of others through a ‘Social Decision

Making Meeting’ (Glick & Gibbs, 2011, p.81). Here individuals must justify the

decisions made in response to questions surrounding vignettes, which are

chosen according to their suitability to the group and complement the skills

learned previously (Glick & Gibbs, 2011). Peers with more mature moral

reasoning challenge their decisions in the hope that the inner conflict

experienced may stimulate more mature understanding (Gibbs, 2004). Dukerich

et al. (1990) found support for this method. Groups achieved lower levels of

moral reasoning when the less principled individuals took a leadership role,

whilst groups with more principled leaders received higher individual scores

following the sessions.

34

2.4.2.2 Cognitive Behaviour Therapy

The theoretical basis of ART is CBT. Each component of the intervention

focuses on a different element of the CBT process, with ACT being concerned

with emotion regulation, Moral Reasoning focusing on developing mature

thought and Social Skill Training concentrating on modifying behaviour

perceived as aggressive (Glick & Gibbs, 2011).

2.4.2.2.1 Definition of Cognitive Behaviour Therapy

CBT is a psychotherapeutic method which aims to support the client in

identifying and modifying internal psychological states, including distorted

thought processes and corresponding negative emotions, in order to instigate

associated changes in a person’s external behaviour. When external

reinforcement in response to this change is triggered, the new behaviours are

maintained (Hollin, 2004). CBT does not describe a single method, but

generally the CBT process consists of several phases including assessment, to

identify the problem and set goals; intervention or psychoeducation, to apply

cognitive and behavioural change techniques; and review, to evaluate the

strategy’s success (Fuggle, Dunsmuir & Curry, 2013).

Others can support this maintenance of new skills by becoming co-therapists

and agents of reinforcement or changing their own behaviours to avoid

encouraging relapse outside of the therapy sessions (Telford & Farrington,

1996). The effects of child therapy tend to be more powerful when key adults

are involved (Dunsmuir & Iyadurai, 2007).

2.4.2.2.2 Origins and Underpinnings of Cognitive Behaviour Therapy

CBT arose from dissatisfaction with the previously dominant approach,

Behavioural Therapy (Westbrook, Kennerley & Kirk, 2011). Reinforcement

interventions did not produce enduring, generalisable changes in behaviour and

negatively influenced internal motivation (Hughes, 1988).

CBT is consistent with a constructivist approach. As individuals interact with the

environment, they construct schemas which allow them to interpret events.

Identification and modification of these schema is a vital part of CBT (Marshall,

1996). These ideas are reflected in early cognitive therapy. For example, Beck

(1991) described that depression stemmed from ‘...the patient’s tendency to

35

interpret his experiences in terms of being deprived, deficient or defeated.’

(p.82). Meichenbaum (1979) and Ellis (1975), two prominent psychologists in

the area of CBT, also argue that concepts and beliefs impact upon our

interpretation of external events. Ellis (1975) stated that man ‘....can rid himself

of most of his emotional or mental unhappiness...if he learns to maximize his

rational and minimize his irrational thinking’ (p.36).

2.4.2.2.3 Applications of Cognitive Behaviour Therapy

CBT has been implemented to support people with a range of needs, from

depression and anxiety (Southam-Gerow & Kendall, 2000), to physical health

problems and relationship difficulties (Marshall & Turnbull, 1996).

Individuals considered to show aggressive behaviour have frequently been

found to experience distorted social cognitions (Lochman & Dodge, 1994; Zelli

et al. 1999). For example, Beck (1999) observed that thoughts of being

wronged led to feelings of anger and a longing for retaliation. These thoughts

were often disproportionate or wrongly attributed, but could be modified when

questioned and evaluated. Strong support is found in the literature in regards to

the use of CBT for aggression (Hollin, 2004). It is now one of the most common

forms of anger management therapy and appears to be effective in alleviating

such issues (Beck & Fernandez, 1998).

CBT is a promising method for use with children and young people (Fuggle,

Dunsmuir & Curry, 2013; Hughes, 1988). Positive results have been gathered

from studies implementing CBT in school settings, improving self-control of

behaviour considered aggressive (Krishnan, See Yeo & Cheng, 2012; Squires,

2001). Skills training and multimodal interventions received the greatest effect

sizes (Sukhodolsky, Kassinove & Gorman, 2004) suggesting that these forms of

CBT are the most effective for reducing such behaviour. Further research

evaluating CBT interventions within UK school environments is still needed

(Rait, Monsen & Squires, 2010).

Cognitive techniques require that an individual has the ability to think

introspectively and reflect on their own thoughts and feelings (Beck, 1991),

which may be an issue with younger children. CBT also appears to be more

beneficial for those with more advanced levels of cognitive development

36

(Durlak, Fuhrman & Lampman, 1991). As children are often referred for CBT,

they may not recognise that they need to change or be motivated to take part in

the therapy (Stallard, 2007). Finally, treatment focusing solely on the child may

not be successful if contextual influences such as family factors are not also

receiving support (Southam-Gerow & Kendall, 2000). Active parental

involvement has been found to increase the beneficial effects of CBT

interventions (Sofronoff, Attwood & Hinton, 2005).

2.4.2.3 Aggression

2.4.2.3.1 What is Aggression?

Aggression is a form of anti-social behaviour (Clarke, 2003), often arising from

anger caused by misperceptions of the social world (Glick & Gibbs, 2011). It

can be defined as ‘...any form of behaviour intended to harm or injure someone

against his or her wishes’ (Breakwell, 1999, p.9), which can be physical or

psychological (Breakwell, 1999). Aggressive behaviour in schools has been

described as being a continuum, from behaviours considered to be low level

acts such as disruptiveness, moderate behaviours such as bullying and

behaviours deemed more serious such as physical fights and group aggression

(Goldstein et al. 1995).

2.4.2.3.2 Cause of Aggression

Aggression may be underpinned by a complex range of causal factors (Pepler &

Rubin, 1991). Biological causes, such as brain activity (Harris, 1978), and

congenital factors (Brennan, Mednick & Kandel, 1991) have been implicated in

the development of aggressive behaviour. However, such studies often

employed flawed methodology (Harris, 1978).

Some believe that the role of learning is more significant (Clarke, 2003;

Goldstein, 1999). For example, research suggests that antisocial and

aggressive behaviour is associated with child rearing practices such as poor

parental supervision and inappropriate disciplining (Farrington, 1995; Kratcoski,

1985). Peers can also act as models and provide reinforcement of aggressive

behaviour, as children perceived as aggressive tend to socialise with others

who are viewed in a similar way (Cairns et al. 1988).

37

In order to provide a concise, coherent literature review the following section

largely focuses upon the interactionist approach to aggression which aligns

closely with the ART intervention. Theorists advocate that a combination of

many factors is likely to contribute to aggression (Rubin, Bream & Rose-

Krasnor, 1991), hence the selection of a theoretical perspective which considers

both internal and external factors. Other texts offer broader description of the

different theories and origins of aggression (Bandura, 1973; Clarke, 2003;

Hersov, Berger & Shaffer, 1978; Pepler & Rubin, 1991).

The interactionist approach has a basis in Lewin’s (1936) formula B = f(p,e),

where behaviour is viewed as a function of both the person and the

environment. Social learning theory (Bandura, 1969, 1973, 1977) was a

particularly prominent interactionist theory. Bandura (1969, 1973, 1977)

described that external conditions, such as the possibility of reward, interact

with the individual’s learning history, an internal feature consisting of information

gleaned from the observation of models or ‘...observational learning...’

(Bandura, 1977, p.12).

Opportunities to observe models carrying out aggressive behaviours and the

consequences then received for performing such behaviours influence whether

aggressive behaviours are learnt and repeatedly acted out (Bandura, 1973;

Bandura, Ross & Ross, 1963). Bandura (1977) believed that most learning

occurred vicariously, by watching someone else perform the behaviour and the

consequences they receive the child learnt which behaviours to imitate. Once

learning has occurred, when confronted with aversive stimuli, the emotional

arousal experienced can lead to a range of behaviours, including aggression,

depending on which behaviours the child has seen used in such situations

(Bandura, 1973).

Social learning theory can also be applied to modify and control aggression.

The same principles apply as mentioned previously, individuals considered to

show aggression observe role models performing more desirable behaviours

and being rewarded in way which the observer finds appealing (Bandura, 1973).

This process is most effective if the individual has the opportunity to watch

38

several different models and is then rewarded for applying such socially skilled

behaviours themselves.

Social learning theory is not without criticism. Whilst it has many practical

applications and can explain the cultural differences found in aggression it

arguably underestimates the active role of the learner. It also cannot explain

instances where behaviour, which is punished, is still maintained (Breakwell,

1999) and does not consider the person’s intent to harm, which is key to the

concept of aggression.

2.4.2.3.3 Outcomes Associated with Aggressive behaviour

Research has found that aggression is fairly stable over time without

intervention (Crick, 1996; Farrington, 1995; Huesmann et al.1984; Kokko &

Pulkkinen, 2005). Negative outcomes associated with early aggression include

social maladjustment and peer rejection (Crick, 1996); anti-social or criminal

behaviour (Farrington, 1995; Huesmann et al.1984; Stattin & Magnusson,

1989); lowered engagement with education and academic achievement

(Gutman & Vorhaus, 2012) and lowered occupational prestige and health

concerns (Huesmann, Dubow & Boxer, 2009). In order to avoid such outcomes

later in life schools are seen as playing a vital role in providing early intervention

and educating pupils in alternative behaviour (Gable, Bullock & Harader, 1995).

As Goldstein (1999, p.2) states ‘Catch it low to prevent it high....’.

2.4.2.3.4 Approaches that are Effective in Reducing Behaviour Perceived to be

Aggressive

Guidance from the Home Office suggests that aggression and violence in

schools is best addressed using approaches such as CBT, social skills training

and mentoring, which have been evaluated extensively (Home Office, 2013).

According to a recent review of meta-analyses, small group skills training in

social competence and anger management, appeared to be highly effective at

reducing anti-social behaviour (Ross et al. 2011).

Effective programmes share several further underlying factors; they are often

multidimensional, underpinned by the principles of Social Learning Theory and

are flexible and dynamic in their approach, tailoring the programme to each

individual (Caldwell & Van Rybroek, 2013). Structures within the school also

39

ensure success, for example school policies which are practised consistently

and supported by well-trained staff (Brown & Winterton, 2010). Ross et al.

(2011) noted the lack of research from the United Kingdom, concluding that

there is a need for more robust, high quality evaluations of promising

programmes in the UK.

2.4.2.4 Social Competence

In contrast to Glick and Gibbs’ (2011) ART, which views the reduction of

aggression as its primary goal, the European version of ART focuses more on

the programme’s ability to nurture skills of social competence (Gundersen et al.

2014).

2.4.2.4.1 What is Social Competence?

Social competence is sometimes referred to as ‘knowing what to do and how to

do it’ during social interactions (Frederickson & Cline, 2009. p.460). This

definition implies that both behavioural elements, such as enactment of

appropriate social skills, as well as cognitive aspects, such as social perception

and problem solving ability, contribute to competent social interaction. However,

definition is problematic because of the many skills associated with social

competence (Vaughn & Waters, 1981). Argyle (1994) poses that social

competence has six components: Social skills and techniques; rewardingness;

empathy; social intelligence and problem solving; assertiveness; verbal and

non-verbal skills. Also, as a social concept, what is seen as appropriate or

desirable depends on the society or culture in which the interaction is occurring

(Frederickson & Cline, 2009).

Social Information Processing Models, such as that constructed by Crick and

Dodge (1994), describe social competence as a cognitive and behavioural

phenomenon which requires several types of skilful processing, including

encoding, interpretation and amalgamation of social information, as well as

response construction, evaluation and finally performance. Difficulty at any of

the stages will result in problems with social relationships. Research evidence

supports the model by finding different patterns of processing in young people

displaying behaviours considered to be pro- and anti- social. For example,

those who act pro-socially view aggressive responses negatively and are less

40

likely to consider others actions to be hostile in intent (Nelson & Crick, 1999),

whilst young people perceived as aggressive appear to experience opposing

patterns of thought (Lochman & Dodge, 1994; Zelli et al. 1999).

2.4.2.4.2 Development of Social Competence

Attachments and early social experiences with caregivers are considered to lay

the foundation for the development of social understanding (Semrud-Clikeman,

2007). It is through these close relationships that the child learns about the

social world, creating internal working models which provide a basis for social

interaction (Bowlby, 1969) as well as modelling behaviour (Cartledge & Milburn,

1995). Parents can influence these skills of social competence directly, such as

supervising them when playing with others, or indirectly, through attachment

behaviours (Ladd, 1999). Rose Krasnor et al. (1996) found a positive

relationship between security of attachment and social engagement in much

younger children. They also found that parenting practices such as maternal

directiveness, were associated with increased social behaviours considered to

be aggressive. Good quality parent-child relationships have also been found to

relate to increased social skill performance later in life, such as during middle

adolescence (Engels et al. 2001). Interactions with peers also play an important

role in the development of social competence (Hay, Payne and Chadwick,

2004).

2.4.2.4.3 Social Competence and Aggression

The Social-Information Processing Model (Crick & Dodge,1994) hypothesised

that an inability to solve social problems, because of inappropriate or lacking

strategies, led to faulty processing of information. This would then trigger

aggressive behaviour. Distorted processing can occur at a number of stages

during the interaction, from the interpretation of cues, to the selection of

appropriate response (Crick & Dodge, 1994).

Supporting research has suggested that young people who engage in

behaviours considered to be violent employ a number of distorted social

cognitions from misinterpreting social cues to problem solving deficiencies

(Lochman & Dodge, 1994). Adolescents perceived to be aggressive also apply

more social problem solving strategies involving the use of aggression, when

41

compared to young people who are considered to be more prosocial (Pakaslahti

& Keltikangas-Jarvinen, 1996). Deviant processing, with a tendency to interpret

the actions of others as hostile, is also often found to be a characteristic of

individuals who are believed to be aggressive (Zelli et al. 1999). Further support

stems from evidence suggesting that different forms of aggression, with

different associated goals, are associated with different patterns of maladaptive

social information processing (Crick & Dodge, 1996). In relation to planning

intervention, Rubin, Bream and Rose-Krasnor (1991) found that the agonistic

strategies employed by these children were often highly successful, making

them resistant to change.

2.4.2.4.4 Social Competence Training

A meta-analysis of 49 studies found that training in social competence was

moderately effective (Beelman, Pfingsten & Losel, 1994). However, long term

benefits were rare and the programmes appeared to be less effective when

measured on broader constructs. Multimodal programmes were praised by the

authors, because they often led to superior generalisation of skills.

Programmes aiming to improve social competence arise from several different

approaches (Beelmann, Pfingsten & Losel, 1994). Some believe that such

children are lacking in behavioural skills and therefore must be taught through

modelling and reinforcement (Goldstein, 1973). Others focus on improving

cognitive skills or modifying inappropriate thoughts (Beelmann, Pfingsten &

Losel, 1994), also known as social problem solving. Fraser et al. (2005)

describes the results of one such intervention, which aimed to strengthen social

information processing and emotion regulation. The results suggested that

training increased social competence and lowered aggression (Fraser et al.

2005). A combination of different methods, comprising modelling,

reinforcement, social skills tutoring and coaching, tailored to the skills deficits of

the individual child and carried out with the support of peers to improve

generalisation, is advocated as the most promising package (Hops, 1983).

42

2.4.3 The Efficacy of ART

Previous research investigating the efficacy of the multimodal social

competence intervention focused upon in the current research, ART, will now

be considered in more detail.

2.4.3.1 Recent Research in Educational Settings

Several of the published research articles into ART have been conducted in

Norway. Gundersen and Svartdal (2010) recruited 140 participants from a wide

range of schools, from Kindergarten to Intermediate level. Results suggested

that the intervention group experienced significant reductions in behaviour

problems and increased social skills. However, they found that the control

participants also showed improvement on these two variables. The researchers

believed that secondary diffusion may have occurred, with the improved

behaviours displayed by the intervention participants influencing those who

were not in the ART groups. The measures did utilise several sources of

reporting but they did not correlate and parents and teachers provided markedly

different scores on the same subscales. More direct measures may have

improved the validity of the data.

Another large randomised control trial study, employed measures at 4 different

time points in order to establish baseline and follow up data (Langeveld,

Gundersen & Svartdal, 2012). 112 children and adolescents were placed into

18 groups across Norway. Data from the intervention group suggested that both

social competence and problem behaviours had improved at post-test, with

problem behaviours decreasing further at post-post-test, 4 weeks later. Further

analyses concluded that the positive changes in social competence mediated

the intervention’s effect on problem behaviours.

Gender, age and initial levels of social competence also mediated the effects of

the intervention. Whilst girls had higher social competence and lower problem

behaviours prior to the programme, both genders benefitted from the

intervention, showing increased social competence. However, boys

experienced a greater decrease in problem behaviour, which was attributed to a

floor effect in the females. Whilst younger participants (under the age of 12-13

years) benefitted from the intervention on both scales, older participants only

43

experienced a slight positive change in behaviour, not social competence. It is

important to note that the authors themselves report that there may have been

elements of selection bias and high non-response may have impacted upon the

validity of the data.

More recently, Koposov, Gundersen and Svartdal (2014) implemented ART in

10 schools and institutions across North-West Russia. The 232 children were

randomly placed into intervention and control groups. Self-report data

suggested that those in the intervention group experienced a significant

increase in their social skills, compared to controls. However, this was

dependent on age, with younger participants (aged 6-9 years and 10-14 years)

benefitting more than the older young people (15+ years). Parent and teacher-

report found that both the intervention and control participants increased in their

social skills and decreased in their problem behaviours. Similar to Gundersen

and Svartdal (2010), the researchers concluded that these findings were due to

the transfer of changes from the ART groups to the control participants in their

daily interactions. Koposov, Gundersen and Svartdal (2014) did state that the

results may have been affected by some teachers not following the instructions

to randomise participants. There were also implementation issues such as

absence and participant refusal. They advocated for future study to consider the

relationship between implementation quality and the outcomes of ART.

The positive outcomes highlighted above suggest that ART is a promising

intervention, in that it has previously been associated with increased social skills

and reduced problem behaviours. This body of research also implies that the

intervention can be accommodated within educational settings. However,

secondary diffusion and threats to validity do complicate the conclusions that

can be drawn from this research.

2.4.3.2 Research in the UK

Despite being available throughout the probation service in the United Kingdom

(NOMS, 2010) there is very little research evaluating ART in the UK. The

studies described below were conducted on adults, as this is how the English

Justice Service chose to adapt and implement the intervention (NOMS, 2010).

Reoffending rates were 14.1% and 13.3% lower for those who had received the

44

programme compared to control participants (Sugg, 2000, cited by McGuire &

Clark, 2004; Hatcher et al. 2008). Those who did not attend the full programme

also had higher rates of reconviction, suggesting that attending all sessions was

vital to ensure positive outcomes. However, both studies suffered from high

rates of attrition which may suggest that the remaining samples were biased.

2.4.3.3 Qualitative findings

Applying ART in schools meant that teachers became implicated in its

implementation. In order to ascertain teacher’s opinions of ART, Sudbeck

(2010) conducted a qualitative study, interviewing 6 ART trained teachers. All

agreed that ART was useful to students as it gave them skills which applied to

everyday life, with moral reasoning being seen as the most important

component. However, it was seen as more applicable to boys than girls. The

majority were confident in their ability to implement the programme and two

related this to their belief that the programme was effective. Lowered

confidence was associated with student’s lack of investment in ART. Three

teachers felt that ART had cultural biases in its curriculum which may impact

upon the student’s dedication to the programme. Therefore some deviated from

the structure to make it more relevant to the individuals in their groups. Finally,

whilst in most cases students were described as reticent at first, their opinions

became more positive as the sessions progressed.

2.5 Implementation Science

A recently defined area of psychology which aligns well with the research and

application of intervention programmes is that of implementation psychology.

Whilst the development and identification of evidence-based programmes and

practices has improved considerably in recent years, the science concerned

with how to implement these programmes lags behind (Fixsen et al. 2005). It

appears there is a ‘...gap between our knowledge of effective treatments and

services currently being received by consumers.’ (p.2).

Implementation science aims to explore and explain what makes interventions

effective in real world settings (Kelly, 2012). The link between level of

implementation and outcomes is so strong that researchers conclude that all

evaluation research should include implementation data (Durlack & DuPree,

45

2008) and that these measures should incorporate a focus on multiple elements

of the implementation process, from treatment integrity to the quality of the

institutional support (Kam, Greenberg & Walls, 2003).

Variability in implementation can arise from several sources but predominantly

characteristics of practitioners and the surrounding context influence

programme effects (Kelly, 2012). Researchers have identified numerous factors

which are considered important in ensuring successful programme

implementation. According to Fixsen et al. (2005) this occurs when:

Staff are selected carefully and receive training and coaching;

there is regular assessment of the fidelity of the implementation of the

intervention and evaluation of staff skills;

administrative support provides clear leadership and supports the

processes and staff;

systems interventions create an environment which is conducive to the

implementation of the programme, including financial, organisational and

human resources to support the practitioners.

These core components influence organisational culture and staff behaviour

best by being integrated as they compensate for one another when one

particular area is weak (Fixsen, et al 2009). Denton, Vaughan and Fletcher

(2003) suggest that the sustained implementation of high quality reading

interventions depends on several additional factors including teacher’s

acceptance and commitment to the programme. They also highlight

organisational factors as being influential in programme success, including the

intervention being valued by management and supported by facilitative

administration.

Domitrovich et al. (2008) note the large variety of contextual factors that

research has highlighted as having an impact on the quality of implementation

of evidenced-based practices in schools. In an attempt to address the research-

to-practice gap present, a three-level model is presented with the aim of

providing a framework for consideration of these many contextual factors

implicated in the implementation of school-based interventions (Figure 2.1)

46

Fig.2.1. A Multi-Level Model of Factors that Influence Implementation Quality. (Domitrovich et al.

2008. p.8)

At the centre of the model lies the intervention package and its corresponding

support system, for example staff training. These are seen as the core

components which should be monitored to ensure high quality implementation.

This is surrounded by several ‘levels’ of interactive factors, each of which can

have an impact on the quality of the implementation of the intervention, both

directly and by influencing other surrounding factors.

By attending to implementation and intervention practices in research, issues

which impact upon effectiveness can be discriminated and provide vital

feedback for further service development (Fixsen et al. 2005). This is an area

which researchers in ART have suggested would be useful to pursue (Koposov,

Gundersen & Svartdal, 2014).

2.5.1 Implementation Psychology and ART

Whilst research suggests that high quality implementation of interventions leads

to more positive outcomes (Durlack & Dupre, 2008), it has also been suggested

47

that poor implementation may actually lead to negative effects (Felner et al.

2001). For example, Barnoski and Aos (2004) present findings which indicate

that poor implementation of ART could actually lead to increased recidivism in

young offenders.

Gundersen and Svartdal (2005) describe the outcomes of a two-year

programme designed to train facilitators in the implementation of ART in order

to maintain its integrity. Following the first year of training the majority of

students reported feeling competent in their skills of implementing ART and

tests suggested that they had obtained sufficient theoretical qualifications to

conduct ART. They also conducted 12 ART groups, after which group members

showed significantly increased social skills and decreased antisocial

behaviours. This suggests that not only was the training successful in providing

the students with knowledge about the programme, in order to improve their

implementation of ART, but this, in turn, led to successful outcomes for the

group members.

2.6 A Systematic Literature Review of Aggression Replacement

Training

2.6.1 Purpose of the Systematic Literature Review

Systematic reviews offer researchers a method for assimilating relevant

evidence, so that robust conclusions can be made (Hammersley, 2001).

Adopting explicit and transparent criteria in the search for evidence minimises

bias, whilst rigorous quality assessment maintains critical awareness of any

issues of reliability or validity (Gough, 2007). This process provides a reliable

base to inform decision-making and can identify areas for future research

(Petticrew & Roberts, 2006).

ART is promoted as ‘...one of the best-validated programmes in its field.’

(Gundersen et al. 2014, p.6). Evaluation studies investigating the outcomes of

ART have employed participants from a range of age groups, including adult

offenders (Hatcher et al. 2008) and Kindergartners (Gundersen & Svartdal,

2010). Evidence also suggests that ART is being adopted by agencies across

several different countries (NOMS, 2010) and settings, from runaway shelters

48

(Nugent, Bruley & Allen, 1998/1999) and young offender institutions (Holmqvist,

Hill & Lang, 2009) to schools (Gundersen & Svartdal, 2010). Despite this wealth

of research it would appear that a systematic review into ART has not yet been

undertaken. Whilst there is one in press (Kaunitz et al in press), it maintains a

focus on both adolescents and adults.

In order to remain consistent and relevant to the work being conducted in the

Local Authority this search will focus on studies of ART implemented with

adolescent participants. Therefore the aim of the current systematic review is to

consolidate the existing evidence base pertaining to the effects of ART when

used with adolescents. This is in order to gain a greater understanding of the

effectiveness of ART for this age group, become familiar with the different

methods employed in its evaluation and identify gaps in the literature which may

provide focus for the current study.

2.6.2 Research Question for the Systematic Review

The primary question for the following review was:

What impact does the ART intervention programme have on adolescents?

Within this several subsidiary questions were also being explored, specifically:

What contexts has ART previously been evaluated in, in relation to physical

setting and wider geographical environment?

What difficulties have those targeted for the intervention experienced?

Who has been involved in the facilitation of the ART sessions?

What methodology has been employed in such evaluations and does the

current evidence base contain research that is considered high quality?

What types of measures have been employed and what variables do they

measure?

2.6.3 Methods Employed in the Systematic Review

This review adhered to guidance provided by the Evidence for Policy and

Practice Information and Co-ordinating Centre (EPPI, 2007). Reference was

49

also made to several introductory texts (Gough, 2007; Gough, Oliver & Thomas,

2012). A more detailed account of the search process is included in Appendix I.

2.6.3.1 Search Strategy

Four databases, which formed part of the University of Nottingham’s e-library

gateway and a general internet search engine were employed in the search,

which used a variety of search terms, devised with reference to the databases’

thesaurus tool (see Appendix I). Initial screening involved reading the titles and

abstracts in order to ascertain whether the paper was relevant in that they

focused upon the ART programme, were written in English and concerned

adolescents. Duplicates were also removed. This yielded 29 papers in total.

During the second cycle of filtering, full text articles were accessed and

inclusion and exclusion criteria employed.

2.6.3.2 Inclusion and Exclusion Criteria

Table 2.1 provides an overview of the inclusion and exclusion criteria used in

the review

Studies were included if they: Studies were excluded when:

Reported primary empirical data Secondary data was reported

Included a pre-measure Pre-measures were not included

Were published in a peer reviewed journal They were not published in a peer reviewed

journal (i.e. theses)

Were evaluative in nature, reporting

quantitative outcome data

The article was solely descriptive

Employed the original ART programme The intervention employed was a modified

version of ART. For example where the

programme has been adapted for special

populations (Moynahan & Stromgren, 2005),

used in conjunction with other methods

(Holmqvist, Hill & Lang, 2009) or condensed

by taking out entire components (Nugent,

50

Bruley & Allen, 1999)

Focused solely upon the adolescent

population, aged between 10 and 19 years

The sample included children under the age

of 10 and adults over the age of 19

Table 2.1: Inclusion and Exclusion Criteria Employed During the Systematic Review

The first three criteria were employed in the hope of ensuring that the data

came from high quality research, resulting in robust conclusions about the

effects of the intervention. Evaluative research was most appropriate to answer

the review question, which aimed to ascertain the impact of ART. The

programme had to be the 3-component, unmodified version of ART in order

relate to the current project. Finally the age range was based on the purpose of

the review and employed to ensure that the results would be relevant to

educational settings, which is where the current project will be based.

In total five studies were identified. A map of the research was created in order

to ease the synthesis and evaluation of the studies and to support the creation

of detailed summaries (See Appendix II).

2.6.4 Comparison of the Studies

In order to explore the subsidiary questions associated with this review, and in

turn, identify areas for future research the studies were compared on the

following factors: context, sample, methodology, implementation and outcome

measures.

The five studies identified were conducted between 1987 and 2009. Two pieces

of research were conducted in both Australia and the United States of America

and one in Norway. Only one study employed multiple settings and in this

research each setting tended to include both a control and experimental group.

The settings varied widely, with one piece of research including both clinical and

educational settings (Gundersen & Svartdal, 2006), two studies conducting their

research in residential youth custodial settings, one in a residential treatment

centre for behaviour disorders and one focusing on a single school setting.

Sample sizes ranged from five to 65 participants. These participants were aged

between 11 and 18 years. Two studies used only males whilst the other

research included both males and females. The difficulties experienced by the

51

participants selected for the research ranged from committing violent criminal

offences (Currie et al 2009) to displaying significantly disruptive behaviour in

terms of deficits in social skills and anger control (Jones, 1991).

Whilst four of the five studies employed a form of randomisation to determine

their groupings, two of these studies employed a cluster method, using either

pre-existing groups or stratifying the groups based on levels of need and then

allocating them to the conditions. Two studies used three groups, including both

control and comparison conditions. Two employed experimental and control

groups. The final study adopted a one group, repeated measures design. Most

of the studies maintained the prescribed 30 session 10 week, format for the

intervention. However, one study extended this to 50 sessions, and another 24

sessions per group on average. Only two studies reported any integrity

measures, including supervision, observations and facilitator logs. The

facilitators of the groups included trained staff from the setting (2 studies),

student facilitators from a local university (1 study) and the provisional

psychologists for the setting (1 study).

Finally all of the studies included both pre- and post-intervention data collection.

One study used self-report measures only, whilst the other research employed

multiple sources of data, including staff and self-report. One also included

parents as contributors to the data collection. All of the studies measured the

participant’s knowledge or use of social skills. Other outcome variables that

were focused upon include moral reasoning ability (3 studies), self-control (3

studies), behavioural incidents (3 studies), cognitive distortions (2 studies),

behaviour and attention difficulties and behavioural functioning (1 study) and

aggression (1 study). The majority of the studies (3 of the 5) employed the

custom measures devised by Goldstein et al (1987) to ascertain social skill

acquisition and behavioural incidents.

2.6.5 In-depth Description of the Studies Found

Glick B and Goldstein A. (1987). Aggression Replacement Training. Journal of

Counseling and Development, 65, 356–362.

The original pilot study of the ART intervention was conducted at a residential

facility for young offenders in New York State. 60 males, aged between 14 and

52

17 were randomly allocated to intervention, comparison (which consisted of a

session of brief instructional training) or wait-list control conditions. Self-report

measures revealed significantly greater acquisition and generalisation of social

skills in the ART groups, but no change in moral reasoning ability. Staff also

reported a significant decrease in the number and intensity of acting out

behaviours for the ART groups compared to both of the control groups.

Repeated measures analyses, conducted using the wait-list control participants,

also found a decrease in acting out behaviours reported by staff. When

released the participants who had received ART scored superiorly on ratings of

global functioning. Reliability of the results may be effected by experimenter

bias as the study was conducted by the creators of the programme. Also the

majority of the measures which found significant differences were those devised

by the researchers. Similar effects may not be found when more global

measures of the same outcome variables are employed.

Jones, Y. (1991). Aggression Replacement in a High School Setting. Australian

Journal of Guidance and Counselling, 1(1), 81–99.

18 students aged 13-15 years, attending a high school in Australia, were

selected for the study based on staff perceptions of their aggression levels.

Students were randomly allocated to 3 conditions: An intervention group, a

moral reasoning only group and a no treatment control group. Teacher-report

measures found that the children in the ART group achieved the greatest

decrease in acting out behaviours. Both the treatment and moral reasoning

groups used significantly more coping skills from pre to post-test and both the

no treatment controls and the ART group improved significantly on a measure

of self-control and impulsivity. Self-report measures revealed that both the

treatment and control groups acquired skills on the social skill acquisition test

(Goldstein et al.1987). The author concluded that moral reasoning may be of

limited value without the other components, providing support for multimodal

interventions. According to the researcher, attendance was poor, particularly to

the moral reasoning sessions, which may have impacted upon the results. The

researcher also suggested that the intervention was not high profile within the

setting and the situations tests required subjective marking (Goldstein et al

1987) which possibly contributed to the inconclusive results.

53

Coleman, M., Pfeiffer, S., & Oakland, T. (1992). Aggression Replacement

Training with Behaviorally Disordered Adolescents. Behavioral Disorders, 18(1),

54–66.

In order to ascertain the effectiveness of ART for adolescents whom the

researchers identified as having behavioural disorders, 39 students from a

residential setting in Texas were selected by staff and randomly assigned to

ART and no treatment control groups. The intervention was provided across 50

sessions, with 2 days per week allocated to homework and practise tasks.

Integrity checks were conducted including daily logs and observations. Five

measures were employed assessing social skills, moral reasoning, self-control

and behaviour incidents. However, only one measure, a self-report created by

Goldstein et al (1987) to ascertain social skill knowledge, showed a significant

improvement for the treatment group above the controls. The researchers

concluded that whilst cognitive gains could be found this did not have an impact

on observable behaviour. It is important to note that high levels of attrition may

have led to a biased sample and low levels of aggression at pre-test may have

affected the outcomes, causing a floor effect. Also, anecdotally the programme

was viewed as an ‘add on’ by the facilitators, who changed often, and boredom

was apparent in the group members, which may have impacted negatively upon

the therapeutic relationships.

Gundersen, K., & Svartdal, F. (2006). Aggression Replacement Training in

Norway: Outcome evaluation of 11 Norwegian student projects. Scandinavian

Journal of Educational Research, 50(1), 63–81.

Across various settings including schools, special behavioural schools and a

psychiatric clinic, 65 11-17 year old participants were screened to ascertain the

level of behavioural problems they experienced. Groups of 6 were then devised

to include different levels of need. The aim was then to randomise the groups to

experimental and control conditions, although the researchers state that this

was not always possible. 47 received ART whilst 18 acted as controls. The

intervention was conducted across 13 weeks with an average of 24 sessions

per group, moral reasoning being the shortest (4.8 sessions). Integrity checks

included supervision and questionnaires. Parent measures revealed significant

54

increases in social skills for the ART group. Measures of problem behaviours

also found significant decreases in behaviour difficulties for the ART group

across raters (parents and teachers). However, a self-report measure of

behavioural functioning showed that both the control and experimental groups

decreased significantly. Positive changes for both groups were also gained on a

measure of cognitive distortions. Finally self-report custom measures found

significant increases in social skills and decreases in problem behaviours for the

ART groups that were not present in the controls. Several issues in the design

and reporting may have lead to questionable validity. Firstly only the 3 children

in each group with the most behavioural difficulties were assessed to lighten the

work-load. Diffusion may have occurred between the 5 groups in the same

settings as the controls and small, uneven, non-randomised samples may have

biased the results.

Currie, M. R, Wood, C. E., Williams, B., & Bates, G. W. (2009). Aggression

Replacement Training in Australia: Youth Justice Pilot Study. Psychiatry,

Psychology and Law, 16(3), 413–426.

Five Australian males, aged 17-18 years, attending a youth justice custodial

were referred onto the ART programme by their health workers. The 10 week

programme was conducted by 1 trained facilitator (the provisional psychologist)

and a supporting social worker colleague. Pre and post-intervention self-report

measures revealed significant reductions in aggression (p=0.06) and increases

in self-control (p=0.03). However, no changes were noted on the measure of

cognitive distortions, despite the behavioural data suggesting that changes in

cognitive appraisal of anger provoking situations had occurred. The researchers

concluded that the effectiveness of the intervention may stem from the

interaction of the components, whereby changes in one element generalise

across the others, which therefore cannot always be measured individually.

However, all results should be viewed with caution as threats to internal validity

may be present as there was no control group impacting upon the results, also

self-report measures were the only source of data, which were analysed using a

liberal level of significance (p=0.10).

55

2.6.6 Quality Assessment

In accordance with the guidance published by the EPPI (2007), the Weight of

Evidence Model (Gough 2007) was adopted to systematically review the quality

of the studies, in terms of their trustworthiness and relevance. The purpose was

to ascertain the relative value of their findings prior to conclusions being drawn

(see Table 2.2 for a summary of the weight of evidence of the studies and

Appendix III for the Weight of Evidence criteria employed).

A-The quality of the

methodology

B-The relevance of

the methodology for answering

the review question

C-The relevance of the evidence for answering

the review question

D-Overall Weight of Evidence

Glick &

Goldstein (1987)

Medium High High High

Jones (1991) Medium High High High

Coleman,

Pfeiffer &

Oakland, (1992)

High Medium High High

Gundersen &

Svartdal (2006)

Medium Medium High Medium

Currie et al.

(2009)

Medium Low High Medium

Table 2.2: Summary of the Results of the Quality Assessment of Studies Identified by the

Systematic Literature Review

2.6.7 Final Summary of the Findings

2.6.7.1 Summary of the Quality Assessment

In regard to the quality of the methodology, the majority of the studies

maintained the prescribed format for the ART programme. However, as only

three of the four studies describe integrity procedures, it is unclear whether all

56

of the conclusions stem from valid applications of the intervention. Only one

study (Glick & Goldstein, 1987) included data collection at follow up. Finally,

although most research used multiple measures from multiple sources,

including teachers and self-report, only one study (Gundersen & Svartdal, 2006)

gathered parental responses. Furthermore, only one study (Currie et al 2009)

ensured that all of their measures were valid and reliable. Three of the other

studies employed measures which were provided with the programme

(situations tests by Goldstein et al. 1987), closely aligned with its contents.

Four of the studies used a form of randomisation to allocate their participants to

the conditions and only one study did not employ a control or comparison group

in order to remove the possibility of issues with internal validity.

Due to the open review question, which was employed purposefully so as to

explore the range of research which has been conducted into ART, all of the

studies gathered were considered relevant to the research question. They

contained participants of the correct age groups, intervention programmes were

close to that prescribed by the handbook and quantitative measures of all three

components of the intervention were used.

2.6.7.2 Impact of Aggression Replacement Training Interventions

Overall the research appears to suggest that ART is an effective intervention for

adolescents with behavioural difficulties. All of the studies concluded that ART

improved the group member’s social skills. Three of the five studies also

provided evidence which suggested that ART was successful in reducing

problem behaviours, with one study utilising measures which revealed that the

intervention was successful at specifically reducing aggression.

Although all of the studies included a measure of the moral reasoning

component, positive results were few and not one found that the intervention

group experienced gains that were superior to the control group, although one

study noted that participants in all conditions improved significantly (Gundersen

& Svartdal, 2006). Gains for both treatment and control groups were also noted

for problem behaviours (Gundersen & Svartdal, 2006), self-control and

impulsivity (Jones, 1991), which could be attributed to the ‘secondary diffusion’

57

effect described previously by Gundersen and Svartdal (2010) and Koposov,

Gundersen and Svartdal (2014).

Positive effects were observed when the intervention was applied across

different settings and geographical locations, when the data was gathered from

several different sources and whether the intervention was conducted as

prescribed or shortened. Participants experienced benefits despite experiencing

different forms of problem behaviour, from criminal behaviour (Currie et al 2009)

to diagnosed behaviour disorders (Coleman, Pfeiffer & Oakland, 1992) and

significantly disruptive behaviour (Jones, 1991).

It is important to note that, in three of the instances where significant changes

were reported, they were gathered utilising invalidated measures that were

closely aligned with the programme. Studies which were considered to be of

high quality only found changes on the custom measures devised by Goldstein

et al (1987), despite using multiple validated measures. This could suggest that,

whilst ART does have a significant impact in changing the specific skills

practised as part of the intervention, the intervention does not support the young

people in generalising this knowledge in the execution of a wider range of social

behaviour. Only two studies reported significant improvements for the treatment

groups which stemmed from independent, valid measures (Currie et al. 2009;

Gundersen & Svartdal, 2006).

Conclusions made here are tentative, given the small number of studies which

contributed to the review. The systematic review process is also not without

fault (Hammersley 2001). For example, publication bias will have impacted

upon the studies gathered in this review, as often only positive results are

published. However, this issue is difficult to overcome (Brunton, Stansfield &

Thomas 2012).

2.6.7.3 Identification of areas for Future Study

This review has revealed several gaps in the literature which could provide

direction for future research. Firstly the small number of studies gathered in the

review, despite the broad review question, suggests that further research is

required into the impact of ART when aimed at the adolescent population.

58

Only one study has focused upon the impact of ART when conducted within a

mainstream school setting (Jones, 1991). It would appear that clinical or

criminal samples are most commonly the focus of the research, which could be

considered a reactive approach. Further research focusing on the

implementation of ART in educational settings, with students displaying

behaviour considered to be disruptive, causing them to be at risk of exclusion,

would not only provide insight into the utility of ART in school settings but would

also explore its application as a preventative intervention.

Several studies have highlighted issues with conducting research into anger

control within a single setting, in that often results are confounded by individuals

from each group influencing the behaviour of one another during their

interactions, leading to secondary diffusion. (Gundersen & Svartdal, 2010;

Taylor at al 2005). Future research could separate the groups to avoid such

diffusion, allowing for the delineation of the impact of participating in the

intervention.

Finally, the vast majority of the research employed facilitators who were

affiliated with the setting, measures that are aligned closely with the programme

and only a single study employed more than two sources of evidence. More

comprehensive data could be gathered from research employing more general,

valid measures of the variables under study, which are completed by multiple

sources who observe the child in different environments. It would also be useful

to explore the effects of such an intervention when implemented by members of

an external support agency.

2.7 Introduction to the Following Evaluation of Aggression

Replacement Training

2.7.1 Rationale for the Research

There is a distinct lack of research into the effectiveness of ART when

implemented in the United Kingdom, despite evidence suggesting that it is a

programme currently used in the rehabilitation of adult offenders through

probation services in England (Sugg, 2000 cited by McGuire & Clark, 2004). At

59

the time of writing, there is no existing evidence regarding the effectiveness of

ART when implemented to support young people with behavioural difficulties

attending educational settings the United Kingdom, despite the growing body of

evidence from other countries which appears to suggest that this programme

shows promise. Therefore the following evaluation of ART will be the first

documented in an educational setting in England.

The literature review also identified several areas for future research focus,

which will be incorporated into the following study. Firstly the intervention will be

conducted using a sample of students attending mainstream educational

provision, in order to explore the utility of the intervention when applied to a

non-clinical or criminal sample, in an educational setting. The programme will

be conducted across several settings, in order to avoid treatment diffusion

between the conditions and multiple sources of data collection will be employed,

using independent, valid and reliable measures. The intervention will be

facilitated by professionals from the EPS who have recently been trained in

ART. Finally, in keeping with the recommendations made by Koposov,

Gundersen and Svartdal (2014), integrity measures will be used in order to

ensure that the newly trained facilitators are implementing the programme as it

was intended.

2.7.2 Research Questions and Hypotheses

Primary Research Questions:

Does participation in ART, implemented in school settings in the UK, support

adolescent participants in improving their use of pro-social behaviours and

decreasing the experience of problem behaviour?

Does ART, implemented in school settings in the UK, contribute to the

development of adolescent participant’s moral reasoning ability?

Hypothesis One:

There will be a statistically significant increase in the pro-social skills displayed

by participants following ART.

60

Null Hypothesis One:

There will be no significant improvement in the pro-social skills displayed by

participants following ART.

Hypothesis Two:

There will be a statistically significant decrease in the problem behaviours

displayed by participants following ART.

Null Hypothesis Two:

There will be no significant improvement in the problem behaviours displayed

by participants following ART

Hypothesis Three:

There will be a statistically significant increase in the moral reasoning maturity

applied by participants following the ART intervention.

Null Hypothesis Three:

There will be no significant improvement in the moral reasoning ability of

participants following the ART intervention.

Subsidiary Research Questions:

In order to inform future applications of the intervention key stakeholders in the

research, specifically the ART training provider and EPS, were interested in

gathering views from those involved in this initial implementation of the

intervention. Therefore a subsidiary research question was added:

What are the views of those involved in the initial pilot of the ART intervention

sessions, in relation to programme implementation, contents and effectiveness?

61

3. Methodology

In order to provide a clear understanding of the current research the following

section discusses several methodological and philosophical stances before

providing a rationale for those employed in the current study. The research is

then described including features such as the sampling procedures employed

and measures adopted. Finally data quality issues such as reliability and validity

and ethical considerations are discussed, in relation to the adaptations made

within the design, in order to achieve the highest quality research possible

within the current context.

3.1 Real World Research

Real world research, conducted in the field as opposed to being laboratory-

based, often focuses upon understanding issues with direct relevance to

people’s lives, or ways of overcoming such issues (Robson, 2011). It is thought

that ‘careful, principled, systematic enquiry’ (Robson, 2011, p.4) is the most

appropriate tool available to fulfil this task.

Research within organisations operating in the real world is not an ‘easy option’

(Gray, 2005, p.2). One example is the education sector, considered messy and

multilayered (Cohen, Manion & Morrison, 2008), which does not lend itself to a

controlled approach to research.

Robson (2011) describes that, in real world research, timing is often beyond the

control of the researcher and actions of the administrators of the programme

may alter the conditions in important ways during implementation. Carrying out

research in natural settings also presents particular practical difficulties such as

a lack of control over extraneous variables and problems in achieving the

random assignment of participants to conditions (Robson, 2011, p.97). Gray

(2005) offers several reasons as to why issues in real world research can arise:

Such environments are complex and the individuals within them busy, making

them difficult to access; key stakeholders may have their own agendas, which

may conflict with those of the researcher and finally, competition and financial

constraints which impact upon the organisation can influence the research.

62

3.1.1 Evidence-Based Practice

Originating in the field of medicine, the drive for evidence-based practice in

education is a relatively recent movement, following criticism about the quality

and relevance of previous educational research efforts (Biesta, 2007).

‘Evidence-based practice...looks to research for evidence about the

effectiveness of interventions.’ (Biesta, 2007, p.7). One of the desirable features

of evidence-based practice is that the outcomes allow clear identification of the

benefits to clients and practitioners (Dunsmuir et al. 2009). Using this

information to inform their actions ensures that professionals are fulfilling an

ethical duty, not only to ensure that these new interventions are better than not

acting (Frederickson, 2002), in other words to avoid doing ‘...more harm than

good...’ (Chalmers, 2003, p.37).

3.1.1.1 Evaluation Research

Key questions about an intervention include ‘does it work?’, ‘when does it

work?’ and ‘for whom does it work?’ (Dunsmuir et al 2009, p.56), to answer

these questions evaluative data is needed. In order to ascertain ‘what works?’,

those advocating evidence-based practice in education advise the use of

experimental methods as the approach able to provide the rigour to gain such

information (Hargreaves, 1999; Lochman, 2000; Stoiber & Kratochwill, 2000).

Some have gone so far as to suggest that practice not based in scientific

evidence is substandard and therefore should be banned (Biesta, 2007).

However, as this often entails real world research, instructional and ecological

conditions in schools make it difficult to control intervention procedures and be

precise in the measurement of outcomes (Stoiber & Waas, 2002).

3.2 Philosophical Stances: Ontology, Epistemology and Associated

Methodology

In social research, a paradigm refers to the beliefs, assumptions, values and

practices shared by a research community (Braun & Clarke, 2014). It provides a

frame of reference from which to organise our observations in research (Babbie,

2010). Since the emergence of the discipline of psychology in the late

nineteenth century there has been disagreement surrounding the most

appropriate ways to theorise and conduct research (Braun & Clarke, 2014).

63

It has been suggested that the differences between paradigms can be

summarised into three factors, the ontology, or the beliefs surrounding the form

and nature of reality; the epistemology, or the relationship between the inquirer

and what can be known and finally the methodology, or the ways in which the

inquirer can find out about reality (Guba & Lincoln, 1994). Research paradigms

are numerous and their classification depends greatly on who is deciding which

criteria to focus on (Morgan, 2007). Here four of the paradigms adopted in

social research are outlined in relation to the three factors mentioned previously

(see Table 3.1 for a summary) and the philosophical stance of the current

research indicated.

3.2.1. The Positivist Paradigm

Whilst some view this approach as a ‘...living faith...’ (Schrag, 1992, p.5) other

academic circles view those who still advocate such a paradigm as ‘...naive

science worshippers...’ (p.5). Dominant from the 1930s to the 1960s, the

positivist approach poses that an external objective reality exists, independent

of the individual, which can become known though scientific observation

(Babbie, 2010). These ‘facts’ are used to test hypotheses and are considered

value-free (Robson, 2011). When attempting to discover ‘what works?’ in

educational settings, the positivist paradigm and the drive to develop causal

hypotheses is hard to avoid (Schrag, 1992). However, critics argue that the

paradigm is reductionist, with limited application in the real world. Research

techniques associated with the positivist paradigm have also been criticised for

stripping the context of random variables which may impact upon findings, so

that the results lack external validity (Guba & Lincoln, 1994).

3.2.2. The Constructivist Paradigm

In stark contrast to the positivist paradigm, the constructivist approach poses

that ‘facts’ do not exist externally, but meaning is constructed through

interaction with the surrounding world, which means that different individuals

can construe the same experience in different ways (Gray, 2005). This

approach believes that the researcher and participant influence one another

through their interactions and therefore constructions can change over the

course of the study (Mertens, 2010). Primarily, the task of a constructivist

researcher is to understand others constructions of knowledge and meaning,

64

which often requires the use of qualitative methods, such as interviews and

observations, in order to gain insight into multiple perspectives (Robson, 2011).

3.2.3 Pragmatism

Those following a pragmatic stance tend to be guided by practical matters as

opposed to theoretical underpinnings (Robson, 2011). Pragmatism recognises

the existence of an external physical world, as well as social and psychological

worlds. It places internal individual experience of the external world in action in

high regard (Johnson & Onwuegbuzie, 2004). Knowledge is viewed as both

constructed and based on the external reality which we live in and experience

(Robson, 2011), leading to layers of both single and multiple realities which can

be explored through empirical inquiry (Feilzer, 2010).

Pragmatic research aims to provide data which is useful, hence the suggestion

that such an approach would be appropriate for real world researchers engaged

in problem solving (Robson, 2011). The main focus of pragmatic research is on

human inquiry, or actions undertaken as we experience our day to day lives.

Some pose this as a criticism, as such research is more likely to lead to

incremental changes, rather than revolution change for society (Johnson &

Onwuegbuzie, 2004). In keeping with these beliefs, pragmatists use whatever

methodology works best to answer the questions posed, hence the suggestion

that it makes a good ‘...philosophical partner...’ (Johnson & Onwuegbuzie, 2004.

p.14) for mixed methods methodology, which integrates the findings from

qualitative and quantitative measures.

3.2.4 The Post-Positivist Paradigm

Dissatisfaction with positivism throughout the 1950s and 60s gave rise to post-

positivism (Tashakkori & Teddlie, 2003). Post-positivists accept that

unobservable phenomena (i.e. thoughts, feelings) are still important elements of

human experience (Mertens, 2010) and that research is influenced by the

values of the researcher, as understandings of reality are constructed

(Tashakkori & Teddlie, 2003). However, by recognising and analysing the

possible effects of such biases on the conclusions drawn, post-positivists

attempt to maintain a commitment to objectivity (Robson, 2011).

65

Because of these limitations, researchers accept that the external reality can

only be known imperfectly (Robson, 2011). Therefore, the approach questions

whether researchers can uncover generalisable laws relating to human

behaviour and instead work towards probabilities about reality rather than

certainties (Mertens, 2010). To overcome some of the criticisms of the positivist

stance, the scientific inquiry endorsed takes place in more naturalistic settings,

involving the collection of situational data and emic viewpoints to determine the

meanings that the participants ascribe to their behaviour (Guba & Lincoln,

1994). These uncontrolled, real world settings often require more quasi-

experimental methodology, as elements of scientific experimentation such as

randomisation of participants is impractical (Mertens, 2010).

66

Table 3.1 Summary of 4 Research Paradigms: Comparison of Ontology, Epistemology and Methodology. Adapted from Fien, (2002); Guba &

Lincoln (1994. p.109) and Mertens, (2010)

Factor Positivism Post-positivism Constructivism Pragmatism

Ontology

Realism: an external

reality exists

Critical realism: reality

exists but can only be

known imperfectly

Relativism: realities are

mental constructions

dependent on the

individual

There is a single reality

and there are multiple

interpretations of reality

Epistemology Objectivist: findings

are true and reality can

be studied without

influencing it

Modified objectivist:

findings are probably

true but bias can affect

findings if not controlled

for.

Subjectivist/

transactional: findings

are created through

interaction

Relationships are

determined by the things

that the researcher

decides are appropriate

to the study

Methodology

Experimental. Mainly

Quantitative.

Quasi experimental.

May include qualitative.

Dialectical/

interpretational.

Qualitative.

Methods matched to the

purposes and questions

of the research. Mixed

methods can be

appropriate.

67

3.2.5 Epistemological Stance of the Current Research

Research is not only a technical undertaking, it is also a ‘...personal, ethical and

political enterprise...’ (Fien, 2002. p.145). The values of the researcher and the

institutions involved in the research have a strong influence over the topics and

methods chosen. This study was underpinned by the post-positivist paradigm,

which is in keeping with previous research projects undertaken by the

researcher and their ‘worldview’ regarding the types of data which would

provide the most rigorous programme evaluation. It is also the most common

paradigm adopted in the evidence-based practice movement (Robson, 2011) as

illustrated by the previous research into ART which is predominantly

experimental. The post-positivist stance of the current research was reflected in

the quasi-experimental design, providing quantitative data which will be subject

to statistical analysis.

It has been suggested that social researchers need not align themselves with

any single paradigm, but can use valuable aspects of several to compensate for

the weaknesses of each other (Babbie, 2010; Mertens, 2010). Research has

also found that many different methodological approaches were perceived of

equal value to teachers in relation to their thinking about teaching, with

narratives provoking thought and experiments offering ideas which could be

applied to their own practice (Kennedy, 1999). Whilst maintaining a

predominantly post-positivist approach, the researcher also places value on the

pragmatic paradigm’s advocacy of utilising the methodological approach which

works best to answer the research question posed (Johnson & Onwuegbuzie,

2004; Robson, 2011), hence the use of supplementary qualitative data. Some

would even suggest that such multi-strategy research aligns most comfortably

within post-positivist epistemology (Giddings, 2006).

Additional qualitative data can improve both the relevance and applicability of

findings by providing an insight into the behaviour observed in the quantitative

data (Guba & Lincoln, 1994). In the current research the qualitative data

gathering was employed to explore a subsidiary research question. This aims to

fulfil the goals of the EPS in which the researcher conducted this study, by

providing an insight into the views of those involved in this initial pilot of the

68

programme, specifically in relation to factors that were felt to be effective or may

require improvement, so as to inform future practice.

3.3 The Research Project

3.3.1 Design

The current research consisted of a quasi-experimental, non-equivalent control

group design, conducted across six settings, utilising both a pre and post-test.

Three of these settings became the experimental group and the other three, the

wait-list control group.

3.3.1.1 Alternative Designs Considered

Randomised Control Trials (RCT)

RCTs are often portrayed as the experimental method best for establishing

causation (Robson, 2011). Randomisation provides control over threats to

internal validity (Gray, 2005) and therefore RCTs are seen as providing high

quality data (Fox, 2003). This had lead to RCTs being considered one of the

highest methods of evaluation present in the ‘hierarchy of evidence’ (Scott,

Shaw & Joughin, 2001, p.5).

However, it is accepted that it is often not possible to conduct ‘truly’

experimental research in real world contexts (Gray, 2005), as matching large

homogenous samples and comparison groups is untenable in applied field

experiments (Greig 2001), whether due to ethical or practical constraints

(Robson, 2011). RCTs have also been criticised for the level of control placed

on naturally dynamic, evolving situations (Cohen, Manion & Morrison, 2008)

which, in turn, leads to a lack of external validity (Cook, 2002). This may explain

why randomisation is rare in educational research (Cook, 2002). In the current

project, the use of multiple settings meant that randomisation was not possible.

Single-Case Experimental Design (SCED)

SCEDs are a scientific, rigorous method which can provide detailed information

about the pattern and stability of performance via continuous assessment of an

individual (Kazdin 2003a). However several issues with the design, primarily

pertaining to the study focusing upon social behaviour as opposed to some form

69

of measurable academic skill, alongside several inherent threats to internal

validity including history and testing, meant that this design was not considered

suitable for the current research.

Repeated Measures Design

The researcher did consider testing the same participants under both

experimental and control conditions. However the timings of the academic year

would have impacted upon the validity of the measures taken to create the

‘baseline phase’ as this would have occurred over a summer holiday. The small

group size would also have had considerable implications for the statistical

power of the data.

3.3.1.2 Rational Behind the Design Employed

Employing an experimental design was not only in keeping with the

philosophical stance of the researcher but also meant that the current findings

could be compared to previous research into the effectiveness of ART, which all

adopted some form of experimental design. It also accommodated for some of

the constraints of the research, such as the intervention and control groups

belonging to different settings.

3.3.1.2.1 Limitations of Quasi Experimental Designs

In the current research participants could not be randomised into the conditions

because the EPS had decided to conduct the intervention within six different

schools, whereby each school would run just one ART group within the current

academic year. Matching the participants in the control and experimental

groups was also not plausible across settings. Therefore a quasi experimental

design was employed. This makes the interpretation of findings more complex

compared to a ‘true’ experimental design (Robson, 2011) and may make the

results vulnerable to several threats to internal validity, including pre-existing

differences in the settings or groups of participants (Mertens, 2010). However,

standardised instructions, reliable measures and treatment integrity checks

were employed to attenuate such effects (see section 3.7 for a more detailed

discussion of data quality issues).

70

3.3.1.2.2 Improving Quasi Experimental Designs

Recent comparisons of experimental and quasi-experimental findings suggest

that quasi-experimental designs can reproduce the findings of ‘true’

experiments robustly (Cook, Shadish & Wong, 2008).

There are several ways in which quasi-experimental designs can be improved.

For example, the utilisation of a control group and pre-test does provide some

protection against threats such as maturation, attrition, testing and differential

selection (Mertens, 2010). Also, by testing multiple dependent variables the

researcher can look for predicted patterns of effects which would strengthen the

conclusions drawn (Shadish, Cook & Campbell, 2002). Utilising a quasi-

experimental design also meant that the issues of secondary diffusion found in

previous research (i.e. Gundersen and Svartdal, 2010) would not impact upon

the validity of the findings. Kazdin (2003b) went so far as to suggest that, with

consideration of several design elements, quasi-experimental designs could

support ‘...very strong inferences...’ being drawn (p.169).

3.3.2 Variables

The independent variable was participation in ART, conducted by newly-trained

facilitators, within secondary school settings, over the course of ten weeks.

The dependent variables were the levels of problem behaviours displayed by

the young people and their use of social skills. These variables were measured

both pre and post-intervention, from multiple sources: Parents, teachers and the

young person themselves. An additional dependent variable was the maturity of

the participant’s moral judgements, which was ascertained by analysing their

responses to hypothetical scenarios (please see section 3.6 below for further

detail).

3.4 Context and Participants

3.4.1 Stakeholders

Identifying those with an interest in the outcomes of this implementation of ART

enables greater consideration of the impact of the results. In the present study

the main stakeholders were:

71

The Local Authority: As a new venture which required considerable resources,

the EPS, which the researcher was working with at the time of the project, was

keen to include evaluation research as part of their piloting of the ART

programme. Their interests were not only focused on whether the programme

was effective but also, with further development of the programme planned,

what changes could be made in future implementations to ensure greater

success. Hence the addition of supplementary qualitative measures, to gather

such information.

iCART: As the first formal evaluation of the ART intervention in the UK, the

training providers of ART in Europe, iCART, were interested in the research as

it could possibly encourage further training across organisations in the UK and

the information gathered could inform further development of the programme.

The University of Nottingham: The importance of evidence based practice

was highlighted as part of the training for a Doctorate in Applied Educational

Psychology. In turn the researcher received a great deal of support in

developing skills of programme evaluation.

The schools participating in the research: Schools dedicated considerable

time and resources to both the implementation of the intervention and

supporting the research, with the hope of improving the behaviour of some of

their pupils.

The children involved in the research: Pupil’s interests related to the

potential changes in their thought patterns and behaviour and the benefits of

such changes, such as avoiding exclusion from school.

3.4.2 Selection of Schools

The Senior EP leading the programme within the EPS asked EPs from the

Service to nominate schools who were considered to have supportive SEN

departments and children with the type of needs targeted by the ART

intervention (i.e. anger control or social skills difficulties). A recruitment flyer was

circulated, via the EPs who worked in the schools and the contact person at the

schools (usually the SEN Co-ordinator) was asked to complete the attached

application form (see Appendix IV).

72

Once the application forms had been returned six schools were selected by the

lead EP who felt they had young people with relevant needs and structures in

place to support the implementation of the programme. The lead EP allocated a

pair of newly trained facilitators to each school. Several criteria were considered

in this decision, firstly location of the EP’s school allocation, to ensure the

distance would not be inconvenient and secondly the EP’s specialism was

considered in relation to the characteristics of the cohort of youngsters at each

school and their associated difficulties.

The schools were then placed into two groups of three based on the EP’s

workloads, the first group of three schools received the sessions in the Autumn

Term and the second group the following Spring Term. These became the

experimental and wait-list control groups, respectively.

3.4.3 School Characteristics

All of the schools involved in the research were sub-urban secondary schools

with integrated sixth forms. All of the schools were mixed gender mainstream

provision, except for School F, which was a selective school for boys. Key

characteristics of the six schools are displayed in Table 3.2. This data suggests

that the schools in the study were diverse in nature. The implications of this for

the conclusions drawn will be discussed further in Section 5.

School No. of pupils

on role

% GCSE grades

5 A*-C inc.

English and

Maths

% Free School

Meals

% English as

an additional

language

School A 1275 44 61.4 80

School B 1078 62 39.9 50.5

School C 661 56 43 7.4

School D 860 64 75.4 53

School E 1173 54 65.2 88.9

73

School F 757 99 20.3 30

Table 3.2: Key Characteristics of the Schools Involved in the Research. Schools A-C make up

the experimental group, Schools D-F the wait list control group. Data gathered directly from the

school’s databases as well as Ofsted’s (2013) ‘Data Dashboard’.

3.4.4 Sampling

Following the selection of the schools the researcher met with the identified

contact personnel at the schools to provide information about the requirements

of the evaluation and to support the selection of participants. This meeting

involved a discussion about possible candidates for the programme using the

inclusion criteria (outlined below in section 3.4.4.2) which were made explicit on

the information sheet provided to schools (see Appendix V) and the

professional judgements of the contact professional at the school, in relation to

their view of the suitability of the young people for participation in the

intervention. It is acknowledged that such a selection process could be open to

bias, given that professional judgements could be considered to be more

subjective than utilising standardised measures. However unforeseen

circumstances meant that a social skills scale, which had been identified to

support selection, was not completed by the young people. The implication of

this form of sampling frame will be explored further in section 5.

Each school was asked to identify six young people, aged between 11 and 18

years, to take part in the intervention sessions. Because of the considerable

commitment, in terms of staffing and time, required by the programme and the

high chance of participant drop out due to exclusion from school, two of the

experimental schools chose to include a higher number of participants at the

start of the programme (School A included 12 and School B included 8).

3.4.4.1 Use of Role Models

The trainer from iCART advocated a model of delivery in which role models

were used to support the development of skills (Gundersen, Finne & Olsen,

2006), a model which he has employed in his own research (Gundersen &

Svartdal, 2006). Previous social skill intervention research has found this

approach to be effective (Prinz, Blechman & Dumas, 1994) and researchers in

74

this area have warned of the iatrogenic effects of homogenous groupings of

high risk young people (Dishion & Andrews, 1995; Dishion, McCord & Poulin,

1999). Including role models serves several purposes. Not only do the more

competent individuals provide positive models within the sessions but they can

also become sources of social support outside of the intervention (Prinz,

Blechman & Dumas, 1994). In the context of ART including those with greater

social competence and more mature moral reasoning capacities was also

thought to generate rich discussion and a broader range of perspective-taking

opportunities in the moral reasoning sessions.

However, the current research deviates slightly from this model for several

reasons. Firstly it was deemed ethically questionable to include young people,

who would be giving up a considerable amount of time from their studies, if they

would not personally gain from participating. Secondly, as mentioned

previously, a measure of social competence was originally planned which aimed

to support the delineation of ‘role model’ and ‘target’ individuals from a pool of

preselected children. However, unforeseen circumstances in several schools

meant that there was not time to identify a ‘pool’ of students who would be able

to complete the social competence measure prior to the parental consent

needing to be sent out. Therefore during selection the contact professional from

the school was asked to use their professional judgement to select 50% of the

students whom they felt would act as positive role models, in that they had more

advanced levels of social competence and experienced only ‘low level’ social

behaviour difficulties, as well as half of the group being ‘target individuals’ who

displayed more challenging behaviour needs. The implications of this form of

selection process will be discussed in section 5.

3.4.4.2 Inclusion Criteria

Several inclusion criteria were devised to support the selection of the

participants. The ART Handbook (Glick & Gibbs, 2011) offers the specific

criterion that group members display ‘…deficiencies in pro-social skills, anger

control and moral reasoning capacity.’ (p.19). In order to protect the participant

numbers, the validity of the research and to ensure the young people would be

able to access the measures used in data collection the researcher also

ensured that the students selected had good attendance, were not involved in

75

any other behaviour interventions and had the ability to reflect upon their

thoughts and behaviour.

3.4.5 Characteristics of the Participants

Initially 41 young people were involved in the research project, with 23 in the

experimental group and 18 in the wait list control group. Table 3.3 provides a

summary of the participants based in each school setting.

Setting Total members in group with consent to partake in

the research

Gender

M/F

Mean age Age range Diagnoses of SEN

No. receiving

Free School Meals

No. with English as an

Additional Language

School A 12 10/2 13y1m 11y10m-

13y9m

1 pupil with a

diagnosis of

ADHD

6 10

School B 7 4/3 13y3m 12y5m-

14y4m

NONE 1 2

School C 4 2/2 13y4m 12y5m-

14y4m

NONE 3 0

School D 6 6/0 12y4m 12y-13y NONE 6 0

School E 6 4/2 13y4m 12y4m-

14y7m

1 pupil with a

diagnosis of

ADHD

4 6

School F 6 6/0 15y 7m 14y1m-

16y11m

NONE 0 0

Table 3.3: Participant Characteristics by Setting

3.4.5.1 Involvement in Other Interventions

It came to light that two students from the control group, one student from

School E and one from School F, had continued to receive one to one support

for social and emotional needs throughout the 10 week period between the pre

and post-measures. Therefore any data collected from these students will be

removed from the final analyses.

76

3.5 Procedure

3.5.1 Setting up the Project in Schools

Once the schools had selected the individuals who were to take part in the

programme, the EPS sought parental consent, via the contact member of staff

at the school, for the pupil’s participation in the ART sessions. Once this was

obtained, the ART facilitators for each school in the experimental group visited

the schools and met with the young people who would form the ART group to

introduce themselves and describe the upcoming sessions.

Consent for participating in the research was gained from parents, the teachers

who would be completing the measures and the young people themselves in all

six schools. Further consent was sought from the ART facilitators at the three

experimental group schools, as they themselves would be involved in integrity

measures and questionnaires (see Appendices VI-X for the consent forms).

Consent to partake in the evaluation was not obtained from parents prior to the

pre-measures for three students, one at School B and 2 at School C, these

individuals were still able to participate in the intervention sessions.

3.5.2 The Intervention Sessions

The adapted ART programme employed in the current research is a multimodal

social competence training method based in CBT (Gundersen et al. 2014). The

intervention consists of 10 sessions of each of three components: anger control,

prosocial skills and moral reasoning. The intervention is conducted over a 10

week period, with one session from each component conducted each week.

The contents of the sessions, summarised from the recently revised handbook

(Gundersen et al 2014), will now be described in more detail. Please see

Appendix XI for a brief summary regarding the structure of the sessions.

3.5.2.1 Contents of the Components of ART

3.5.2.1.1 Contents of Anger Control Sessions

The topics covered in the anger control sessions follow a set sequence so that

the skills build upon one another as the programme progresses (see Table 3.4).

For example week 2 covers ‘triggers’ in which group members discuss things

which make them angry, in week 3 this is expanded upon by talking about how

77

the young people know when they are angry and what to do to try and reduce

anger. Sessions 8-10 incorporate skills learnt in the prosocial skills sessions, as

alternative behaviours to aggression. The primary activities during the anger

control sessions are to discuss the topic and role play the skills in small groups,

gathering feedback from the other members. Homework is provided in the form

of a ‘hassle log’, which is a self evaluation tool where participants detail real life

events and how they dealt with them using the skills they have learnt.

Week Anger Control Training

1 Introduction

2 External and internal triggers

3 Cues and reducers

4 Reminders

5 Self evaluation

6 Consequences

7 How do I end up in conflict situations?

8 Choice of pro-social skills-solutions

other than aggression

9 As above

10 As above

Table 3.4 Week by Week Session Plan for Anger Control Sessions Taken from Gundersen et

al. (2014. Pp.84)

3.5.2.1.2 Contents of the Prosocial Skills Training Sessions

The topics for the pro-social skills sessions are based on a more prescriptive

model, chosen from a possible 50, based on the needs of the members of the

intervention group (Gundersen et al 2014). However 10 ‘core’ skills are

suggested in the handbook, which complement the topics in the other

components (see Table 3.5 for a list of these skills). These sessions also use

78

discussion and role play as the main forms of activity. Homework involves

writing a journal of how they have applied the skill learnt that week outside of

the session.

Week Prosocial Skill

1 Giving a compliment

2 Starting a conversation

3 Asking for help

4 Being aware of one’s feelings

5 Asking for forgiveness

6 Rewarding yourself

7 Dealing with peer pressure

8 Asking permission

9 Dealing with persuasion

10 Dealing with an accusation

Table 3.5: Week by week session plan for the prosocial skills sessions (taken from Gundersen

et al 2014. Pp.31)

3.5.2.1.3 Contents of the Moral Reasoning Sessions

The moral discussion groups, which take part during each of these sessions,

are also based on a prescriptive model. The moral reasoning dilemmas are

selected from a bank of examples provided in the handbook or created by the

facilitators, based on their relevance to the scenarios encountered by the young

people in their daily lives. The coaches are considered to play a particularly

important, active role during these discussions, encouraging discussion

between the young people in the group and challenging and clarifying when

needed.

79

3.5.2.2 Implementation

Pupils attended the intervention sessions in groups of 6-12. The schools were

asked to provide a designated room where the groups would be held and a

member of staff from the school attended the sessions to support the EP

facilitators. Timetabling of the sessions was negotiated between the school and

the facilitators and where possible was arranged to avoid students missing core

subjects. Each type of session (social skills training, anger control or moral

reasoning) always took part on the same day of the week at the same time to

provide some consistency for the young people.

As part of the sessions the handbook recommends the use of posters, rewards

and games to maintain ownership and motivation among the students

(Gundersen, Finne & Olsen, 2006). Homework assignments also ensured that

the students practised and generalised the skills outside of the intervention

sessions (Gundersen, Finne & Olsen, 2006). The handbook states that ‘ART

must be included as a key factor in the organisation’s activity’ (Gundersen,

Finne & Olsen, 2006, p.56). According to the authors, the most important factor

in the success of ART is that it becomes entrenched within the organisation.

Providing school staff and parents with information about the intervention is

encouraged in the hope that people within the child’s social network will

reinforce the use of the skills outside of the intervention sessions (Gundersen et

al. 2014).

3.5.2.2.1 Facilitators

Facilitators consisted of 11 EPs and one senior behaviour support professional

who had recently received eight days of training from iCART. These individuals

were placed into pairs and allocated to a single group of students within a

particular school. As part of the training qualification they were asked to

complete a full course of the ART intervention in a school setting. This research

aims to evaluate the outcomes of that pilot.

3.5.2.3 Integrity Checks

A review of intervention evaluation research conducted by Gansle (2005) found

that whilst 30% of the studies mentioned the importance of treatment integrity,

only 10% actually conducted some form of integrity checks. This is concerning

80

considering that research evidence suggests that high quality implementation of

interventions leads to improved prevention outcomes (Durlak & Dupre, 2008),

making integrity checks an ‘essential feature of programme evaluations.’

(p.327).

Due to the intervention in the current research being carried out in different

settings by different facilitators the integrity of its implementation was

considered a potential threat to validity (threats to internal validity are discussed

further in section 3.7.1). In order to ascertain the consistency of the

implementation of the ART intervention and reduce the potential threat three

integrity checks were completed by the researcher in each school, one for each

different component. Checklists available from the iCART website (see

Appendix XII) were employed to analyse the content and delivery of the

sessions, in terms of their relation to the programme. To assess the reliability of

the researcher’s judgements, one session per group was also rated by a fellow

trainee EP using the same checklists.

The outcomes of the checklists were reviewed in order to assess the treatment

integrity of the three groups in the experimental condition. When the three

sessions were added together the checklists contained 93 items, equivalent to a

total of 93 points. Table 3.6 presents the number of items identified along with

the overall percentage fidelity and percentage agreement between the

researcher and second rater.

School Anger

control

session

/25

Social skills

training

/39

Moral

reasoning

session

/28

Total

/92

Percentage

fidelity

Percentage

inter-rater

agreement

for 1

session

A 21 36 24 81 88% 100%

B* Xxx 33 Xxx Xxx Xxx Xxx

C 23 32 24 79 86% 100%

Table 3.6: Results of the ART Treatment Integrity Checks.

*School B left the project before all integrity measures were completed.

81

The most common elements missed out of the sessions were due to time

constraints, as the facilitators had to fit into the school’s timetable in their timing

of the sessions, therefore often the skills were only modelled once (rather than

the two times proffered by the checklist) and the group members only had the

opportunity to role play each skill once (as opposed to twice). Other issues

included not handing out or checking homework, providing feedback after the

role plays in an order other than that specified in the handbook and group

members not role playing voluntarily.

3.6 Measures Used

In order to ascertain the outcomes of each of the three intervention

components, two measures were employed before and after the intervention

sessions.

3.6.1 Measure of Problem Behaviours and Social Skills: The Social Skills

Improvement System

The SSIS-RS (Gresham & Elliott, 2008) includes 15 subscales, which together

measure social skills, competing problem behaviours and academic

competence (see Table 3.7 for a summary of the subscales). Three versions

are available, including teacher, parent and student forms, suitable for two age

groups; 8-12 and 13-18 years. All utilise a 4 point rating scale to indicate how

often children engage in the behaviours described. The manual recommends

10-25 minutes for raters to complete the measures. The current version of the

SSIS-RS was originally standardised on children in the USA, which means that

conclusions drawn from the data collected in the following research, which was

conducted in the UK, should be viewed tentatively. However, the previous

version of the measure has been used in research conducted in other European

countries (Langeveld, Gundersen & Svartdal, 2012) including the UK (Liddle &

Macmillan, 2010; Pritchett et al.2013).

The SSIS-RS has been found to have satisfactory test-retest reliability

(approximately .80 for all three versions) (Crosby, 2011) and high internal

consistency estimates (median coefficient alphas of .90s) have been reported

for the main scales on all three forms (Gresham & Elliott, 2008). However, inter-

rater reliability estimates were lower, with the teacher and parent forms

82

receiving median correlations of .58 and .59 respectively. The authors attribute

this to the scales including infrequent behaviours and one rater in the study

having more prolonged contact with the children than the second rater.

In relation to content validity, the SSIS-RS underwent several processes during

development including utilising criteria from the DSM-IV, factor analysis and

tests of perceived importance by users (Gresham & Elliott, 2008). Internal

correlations between the scales were moderate to high and in the expected

directions (i.e. social skills and problem behaviours were negatively related,

whilst the social skills subscales were positively related). Finally correlations

with similar measures were again moderate to high (Gresham & Elliott, 2008).

For example, correlation coefficients of .44-.98 were gained when comparing

the SSIS-RS with similar subscales of the Behavior Assessment System for

Children-Second edition (Reynolds & Kamphaus, 2007, cited by Gresham &

Elliott, 2008).

Due to the timings of the measures within the academic calendar, teachers and

parents were asked to reflect upon the young person’s behaviour during the two

weeks prior when completing the questionnaires, as opposed to the two month

timeframe provided on the front of the SSIS-RS. This is because the researcher

felt that, as the school holidays immediately followed the end of intervention in

the experimental schools and precluded the start of the intervention sessions in

the autumn, reflections over two months would be confounded by the pupil’s

behaviour during holidays, which is unlikely to be representative of their

behaviour during term-time. Not only this but the intervention itself was only 10

weeks in duration, so reflections over two months at post-measure would

include behaviours observed after only two weeks of input.

Problem Behaviours Subscales Social Skills Subscales

Externalising Communication

Internalising Cooperation

Hyperactivity/inattention Assertion

Bullying Responsibility

83

Autistic Spectrum

(not included in the self report measures)

Empathy

Engagement

Self Control

3.7 Table to Display the Subscales Included in the SSIS-RS

3.6.1.1 Rationale for using the SSIS-RS

Initially the measure was selected based on the recommendation of the trainer

from iCART, who noted that previous versions of the SSIS-RS had been

adopted in other research into ART (Langeveld, Gundersen & Svartdal, 2012).

The subscales also aligned with the skills taught in the ART curriculum and

were organised into wider outcomes, as opposed to focusing solely on direct

measures of each individual skill taught, as in the custom measure from

Goldstein et al (1987). The SSIS-RS was also easy to administer in a group

format and because of the multi-rater forms, the responses could be

triangulated to gather perspectives across home and school environments.

3.6.2 Measure of Moral Reasoning: The Socio-Moral Reflection Measure-Short

Form

The SRM-SF (Gibbs, Basinger & Fuller, 1992) assesses ‘the maturity of moral

judgement’ (p.33). Children rate the importance of 11 items illustrating five

different classifications of moral values such as ‘contract and truth’ or ‘affiliation’

and justify their decision (see Appendix XIII for the full questionnaire). For

example:

1. Think about when you’ve made a promise to a friend of yours. How

important is it for people to keep promises, if they can, to a friend?

Circle one: very important important not important

WHY IS THAT VERY IMPORTANT/IMPORTANT/NOT IMPORTANT

(WHICHEVER ONE YOU CIRCLED)?

(Taken from Gibbs, Basinger & Fuller, 1992 p.150.)

The questionnaire takes approximately 25-40 minutes to complete.

84

A child’s Sociomoral Reflection Maturity Score (or SRMS) is deduced by

calculating the mean average of the questions. At least 7 of the 11 questions in

the questionnaire must achieve a score for the questionnaire to be analysed.

The SRMS reflects the child’s moral judgement stage, for example a score of 3

indicates that they are at the third stage of Gibbs’ moral development (Gibbs,

Basinger & Fuller, 1992).

During the original psychometric investigations Gibbs, Basinger & Fuller (1992)

report having conducted the SRM-SF with subjects as young as 9-10 years of

age through to adults. During this research the questionnaire was group

administered. Acceptable levels of test-retest reliability (correlation coefficient of

.88) and internal consistency (.87) were reported (Basinger, Gibbs & Fuller,

1995).

The manual advises 30 hours of self-training in order to become skilled at

scoring the pupils justifications, which are allocated a moral judgement stage

based on their similarity to a given list of possible ‘criterion justifications’.

Following such self-training the authors found that inter-rater agreement

between novice rater and expert raters is high, with scores of 90.9-100%.

Concurrent validity was established using the Moral Judgement Interview

(Colby & Kohlberg, 1987 cited in Gibbs, Basinger & Fuller, 1992), achieving a

significant correlation coefficient of .69 and construct validity was supported by

a significant difference between scores achieved by different age groups

(Gibbs, Basinger & Fuller, 1992). Whilst this measure was also developed in the

USA, it has been employed in research in over 75 countries (Gibbs et al. 2007).

A study in the UK also suggested acceptable to high levels of validity and

reliability when administered in a group to British 10 and 11 year olds

(Ferguson, McLernon & Cairns, 1994).

Because of the qualitative nature of the data gathered from the SRM-SF and

the possibility of subjectivity affecting the analysis of the results, a fellow EP,

who had recently undertaken doctoral study and had experience in qualitative

methods, inter-rated 25% of the SRM-SF output. Table 3.8 displays the results

of the inter-rater checks in relation to the criteria provided by the handbook

(Gibbs, Basinger & Fuller, 1992). The implications of these findings will be

discussed further in Chapter 5.

85

Inter-rater criteria Handbook specifications Results achieved

Mean absolute

discrepancy in SRMS 20 points 27.2 points

Global stage

agreement within 1

interval

80% 85%

Exact global stage

agreement 50% 38%

Correlation r=.80 r=.62

Table 3.8: Results of SRM-SF Inter-Rater Analyses.

2.6.2.1 Rationale for using the SRM-SF

As opposed to evaluating whether an action is moral or not, as is the case in

recognition measures, it is the maturity of the reasons given to justify such

decisions which are of utmost importance and represent the deeper

understanding of moral values (Glick & Gibbs, 2011). As a production measure

utilising open statements the authors argue that the SRM-SF is more

ecologically valid than measures which provide fabricated dilemmas. The lack

of lengthy dilemmas or extensive lists of multiple choice justifications also

makes the SRM-SF easier for those with reading difficulties (Gibbs, Basinger &

Fuller, 1992). It is also briefer than previous versions, easing group

administration.

3.6.3 Data Collection Procedures

Data was collected at two time points, pre- and post-intervention, in all six

schools. In regard to the pupil measures, at each setting participants met as a

group with the researcher, who first outlined the purpose of the research and

answered any questions about the consent forms. The instructions for each

questionnaire, along with the first question, were read aloud by the researcher.

Both measures were then administered as a group, with a member of staff from

86

the school made available where it was felt that individuals may require support

for literacy difficulties.

The members of staff filling out the teacher-report measures were selected in

collaboration with the contact person at each school, based on them having

regular contact with the target child. The contact person at each setting also

sent out the parent questionnaires, to be completed at home, which were

returned to the schools in sealed envelopes. Where the schools felt that parents

may have difficulties in completing the questionnaires, perhaps due to literacy

difficulties or due to English not being their first language, drop in sessions were

offered by the schools whereby the parents could complete the questionnaires

with the support of a member of staff.

3.6.4 Additional Data

3.6.4.1 Supplementary Qualitative Data

Whilst the evidence-based movement suggests that it is important to know

whether an intervention is successful, it acknowledges that equally important

questions, in complex settings such as schools, include; under what conditions

does the intervention work? And which approaches and strategies support the

outcomes? (Stoiber & Waas, 2002). After all, for real world evaluation research

to be useful, we need to know if the programme worked but also how to

implement and improve it (Blase et al. 2012).

Supplementary qualitative methods were employed in order to explore the pupil

and facilitator’s perceptions following their experience of the ART programme.

This type of descriptive information posed by Stoiber and Waas above was

greatly desired by the EPS as this was a new venture for them. The current

research was acting as a ‘pilot’ to gather data which could be used to improve

future implementation of the programme.

3.6.4.1.1 Qualitative Measures-Pupils

Following the quantitative post-measures, parental and pupil consent were

sought from all of the participants in the experimental groups. Data collection

took place in the form of a one-to-one semi structured interview with the

researcher, with pre-identified areas for questioning including how successful

87

they felt the programme was, what they felt worked well and what could be

changed to improve the experience in future (see Appendix XIV for a copy of

the prompt sheet used by the researcher in the semi-structured interviews).

3.6.4.1.2 Qualitative Measures-Facilitators

Following the completion of the 30 ART sessions for the experimental group, a

questionnaire was sent out to the six facilitators, asking for their views on the

success of the programme, factors which may have impacted upon the

outcomes and changes they would make to the implementation in future (see

Appendix XV for a copy of the questionnaire).

3.6.5 Data Analysis

3.6.5.1 Quantitative

Data relating to Research Questions 1 and 2 was analysed using both

descriptive and inferential statistical analyses, detailed in Section 4. As the data

did not meet the assumptions of parametric statistical tests, non-parametric

alternatives were used to identify any differences between the intervention and

control groups at pre and post-test. Further tests of subgroups were employed

to look for differences between different aged participants and role models

compared to target individuals. Finally non-parametric correlational tests were

used to explore any possibly relationships between test scores and attendance

at the intervention sessions.

3.6.5.2 Qualitative

The interviews with the student participants and the questionnaires completed

by the facilitators were analysed using Thematic Analysis procedures, as

described by Braun and Clarke’s (2006) six stage model, in order to identify

themes which would inform the supplementary research question.

3.7 Data Quality Issues

The aim of experimental research is to provide results which are valid, in that

they show what they intended to show and reliable, so that potentially the

conditions can be replicated and similar results achieved (Field & Hole, 2006).

Here common threats to validity and reliability in experimental research are

88

discussed and measures taken to minimise the effects of these threats in the

current research are described.

3.7.1 Internal Validity

Threats to internal validity impact upon the researcher’s ability to say that the

changes in the dependent variables can be attributed to the independent

variable, in this instance the intervention sessions (Shadish, Cook & Campbell,

2002). Quasi-experimental designs, are particularly vulnerable to such threats

due to the lack of randomisation (Shadish, Cook & Campbell, 2002). Those

commonly referred to in experimental designs and the steps taken in the current

research to minimise their effects are displayed in Table 3.9

89

Threat to internal

validity

Description Actions taken to reduce impact

History Uncontrolled events may occur

during the time that the

research is taking place that

may impact upon the

outcomes.

Information was gathered from the school about any additional input that the children

may be receiving related to social or behavioural difficulties. More than one dependent

variable will be focused upon in the data collection. A wait-list control group was also

used for comparison. However, multiple settings were employed within each condition

so this issue could still be apparent due to differences between the settings, despite

treatment fidelity checks.

Maturation Changes in participants during

the research that are not

related to the intervention.

The time between the two measures was relatively short in length (approx.11 weeks).

A wait list control group was employed. All subjects were roughly the same age (all

adolescents).

Statistical

Regression

The tendency for scores at

post-test to move towards the

mean.

Appropriate statistical analyses will be conducted. Checks will be used to see if the

groups were equivalent at pre-test. The groups are not solely composed of those who

would be expected to achieve ‘extreme scores’ but also role models, who are also

hoped to show improvements at post-test. A no-treatment control group was employed

and therefore effects would also be seen in their results.

Testing/Instrument

Reactivity

Subjecting participants to initial

testing can affect their

behaviour on subsequent tests.

Participants were fully informed of the purpose of the research in line with ethical

guidelines so demand characteristics may have occurred. To maintain consistency all

measures were completed in the same environment, using standardised instructions

and valid and reliable measures were used. Testing only took place on two occasions

and were separated by an 11 week interval. The control group also underwent the

90

testing procedures.

Instrumentation Unreliable measures may

interfere with the data gathered.

Valid and reliable measures were used. Inter-rater reliability was gathered for the more

complex moral reasoning measure. The same member of staff completed the teacher

measure for each participant at both time points. Standardised instructions were read

to all groups.

Selection Selection bias may yield

differences in the groups of

participants which impact upon

their performance.

Originally a screening measure was planned. However time constraints meant that this

was not used. The researcher supported all schools with selection and employed the

same selection criteria. As random allocation was not used and each group belonged

to a different setting it is accepted that uncontrolled pre-existing differences may exist.

Statistical tests will be used to check that the groups were equivalent at pre-test.

However the purpose of the research is not to generalise to other groups/settings.

Experimental

Mortality

Attrition from the original

sample may result in a biased

group.

As the intervention occurred over a relatively short period of time, with the full support

of the schools, it is hoped that attrition will be minimal. Schools were asked to select

‘good attenders’. Schools have also stated that attendance at the sessions will be

mandatory even if the young person is excluded for a fixed period.

Selection-

Maturation

Interaction

The tendency for groups to

move towards each other on a

dependent variable if initially

different

Selection criteria employed for both experimental and control groups. Groups were

similar in age (all adolescents). However the gender ratio was different in the two

conditions.

Diffusion of When control and experimental

groups communicate with each

Although this is an issue in previous research into ART the current research employs

91

treatment other, leading to the passing of

information regarding the

independent variable.

control and experimental groups which are based in different settings.

Table 3.9: Common Threats to Internal Validity and Actions Taken to Minimise Their Effects in the Current Research. Adapted from Cohen, Manion &

Morrison (2007) and Babbie (2010)

92

3.7.2 External Validity

External validity refers to the degree to which the findings of a study can be

generalised to other populations and settings outside of the experiment (Cohen,

Manion & Morrison, 2007). The naturalistic setting of the current research

improves the ecological validity, making it similar to real life instances where the

intervention may be applied in future, in order to generalise the findings.

However the researcher should also be able to demonstrate that the

participants in the study are representative of the wider target population

(Cohen, Manion & Morison, 2007). The use of a small sample in the current

research, which did not use screening measures for identification nor

randomisation to conditions, suggests that attempts to apply the results to the

wider population would be invalid. However, the intention was not to generalise

but to be ‘...interested in a specific finding in its own right...’ (Robson, 2011,

p.91) and provide an indication of the impact of the initial pilot of the intervention

alongside additional qualitative data to guide future practice. It is also hoped

that the findings will also contribute to the existing evidence base and stimulate

further research, as this is the first investigation regarding the effects of the

intervention with a school based population within the United Kingdom.

3.7.3 Reliability

Reliability refers to ‘...the stability or consistency with which we measure

something.’ (Robson, 2011. p.85). Unsystematic errors which vary across data

collection points can result in unreliable data (Mertens, 2010). Several actions

were taken by the researcher to avoid such error:

A script of instructions was created so that participants in all settings

received the same guidance at the beginning of the questionnaires

during both pre and post-test. The SSIS-RS also has standardised

instructions on the front of the form, which the parents and staff were

asked to follow.

Participants in each setting also completed the measures in the same

environment. However, this could not be controlled for the parent and

93

teacher measures, which were sent by post or were hand delivered by

the contact member of staff at the school.

The same member of staff was asked to complete the questionnaire for

each child at both time points.

Valid and reliable measures were employed (see section 3.6).

To ensure that the researcher was consistent in her classifications, inter-

rater scores were collected for the data gathered using the SRM-SF,

which applies a more complex, subjective scoring system, as well as

during the observations of treatment fidelity.

In an attempt to control for participant characteristics (i.e. mood,

motivation, social desirability bias) influencing the results, several

sources of data were used and support was provided for those with

literacy difficulties so that the complexity of the qualitative responding

(the SRM-SF) was not a source of bias. However, it is acknowledged that

these factors may still have impacted upon the results gathered.

3.8 Ethical Considerations

In order to ensure ethical the researcher sought ethical approval from the

University of Nottingham’s Ethics Committee (see Appendix XVI for the letter of

approval received from the committee).

Several ethical considerations pertinent to the current research, in keeping with

the guidance from the British Psychological Society (BPS, 2010), will now be

briefly outlined and the steps taken to deal with such issues described.

Informed Consent

As the participants were under 16 years of age parental consent was gathered

initially. The EPS sent out consent forms for parents to sign, in order to give

permission for their children to participate in the intervention sessions. The

researcher also sent out consent forms via the contact member at the school,

with information sheets attached, providing details about the study. Once these

were returned the pupil’s themselves were asked for consent, as were the

teachers completing the SSIS-RS (see Appendix VI-IX). The pupil information

94

sheet was read aloud at the start of the pre-measures session and the

researcher answered any questions to ensure the details of the project were

clear. The facilitators running the three intervention groups also signed consent

forms detailing the treatment integrity procedures and the post-intervention

questionnaire. Additional consent was gathered from the parents and pupil

participants after the quantitative post-measures for participation in the

additional interviews, which included information about the audio recording

equipment which would be used (Appendix XVII and XVIII).

Right to Withdraw

The right of all participants to withdraw was made explicit on the information

sheets and consent forms. This was reiterated to the young people verbally

before they started the measures. All parties were provided with contact details

of the researcher should they wish to withdraw their data at a later date.

Confidentiality

On all written and audio recorded data the participants were referred to by

coded identifier to maintain anonymity. Participants were ensured of their

privacy with the sole exception of any concern arising relating to them being in

possible danger. All data collected was stored in a locked filing cabinet or on an

encrypted memory stick.

Protection from Potential Harm

As some of the questions on both of the measures could be considered

sensitive, for example the SSIS-RS asks questions about problem behaviours

and the SRM-SF includes potentially upsetting moral dilemmas such as the

death of a friend. Example questions were provided on the parental information

sheets. A member of staff from the school was also made available during the

data collection sessions should any students wish to discuss their concerns.

It was deemed unethical for role models to be young people who displayed no

difficulties in the areas of anger control, moral reasoning or social skills, given

that the intervention would take up a considerable amount of their study time

(30 hours). Therefore the selection criteria was included which stated that role

95

models should have ‘low level’ needs of their own, so that they too would

benefit from taking part, whilst being a positive influence on the ‘target’

individuals.

Finally a wait-list control group was employed so that all pupils identified as

having difficulties would gain support. This was also true of those who did not

return their parental consent for the research in time for the pre-measures, who

were still included in the intervention sessions.

Debriefing

Following the completion of the data analysis procedures all schools, pupil

participants, parents and the facilitators associated with the research project

were provided with a summary sheet of the findings of the research as well as

contact details for the researcher, should they wish to discuss anything further.

96

4. Results

4.1 Quantitative Data Analyses

This chapter presents an overview of the data gathered and analyses

performed following the procedures outlined in Chapter 3. For a more in depth

consideration of the implications of the results please refer to Chapter 5. Initially

details about levels of attrition from the study will be provided. This is followed

by two levels of statistical analysis whereby both descriptive data and inferential

statistical outcomes are displayed.

4.1.1 Dependent Variables and Direction of Change

As described previously in Chapter 3, the study entailed a high number of

dependent variables, mainly due to the subscales associated with the SSIS-RS.

To support the reader in the interpretation of the following statistical analyses

Table 4.1 provides an overview of the dependent variables measured in this

study, as well as the direction of change desired.

Sub-scale scores

Name of

Instrument

Self report

Measures

Parent

Report

Measures

Teacher

Report

Measures

Hypothesised

Direction of

Change

SSIS

Problem

Behaviours

Externalising behaviour

Bullying

Hyperactivity

Internalising behaviour

Problem behaviours

composite score

Problem behaviours

standardised score

Externalising behaviour

Bullying

Hyperactivity

Internalising behaviour

Autism Spectrum

Problem behaviours

composite score

Problem behaviours

standardised score

Externalising behaviour

Bullying

Hyperactivity

Internalising behaviour

Autism Spectrum

Problem behaviours

composite score

Problem behaviours

standardised score

A decrease in

scores

97

SSIS Social

Skills

Communication

Co-operation

Assertion

Responsibility

Empathy

Engagement

Self control

Social skills composite score

Social skills standardised

score

Communication

Co-operation

Assertion

Responsibility

Empathy

Engagement

Self control

Social skills composite score

Social skills standardised

score

Communication

Co-operation

Assertion

Responsibility

Empathy

Engagement

Self control

Social skills composite score

Social skills standardised

score

An increase in

scores

SRM-SF Sociomoral Reflection

Maturity Score reflecting child’s moral judgement

stage

Xxx Xxx An increase in

scores

Table 4.1 Table to Show the Dependent Variables Employed in the Study and the Desired

Direction of Change

4.1.1.1 Power Analyses

Given the large number of comparisons being calculated it would typically be

advisable to conduct post-hoc corrections, such as a Bonferroni correction, to

control for familywise error rate, or the chance of a Type I error, in which an

effect is believed to exist which actually does not. However, such controls lead

to a loss in statistical power, which suggests an increased risk of a Type II error,

whereby an effect that does exist is rejected (Field, 2009).

The researcher conducted post-hoc power analyses using G Power software

(Buchner et al. 2014). Using a medium effect size of d=0.5, based on the

findings of previous research into ART which found effect sizes between 0.35

and 0.63 (Coleman, Pfeiffer & Oakland, 1992; Currie et al. 2012 & Jones,

1991). Given the current participant levels, the analysis suggested that the

following inferential tests would have a statistical power of 0.29, which is the

probability that the tests will find any effect that exists. This suggests a 71%

chance of failing to detect a genuine effect.

98

Apriori analyses also suggested that to achieve a power of .8, a figure

recommended by Cohen (1992) which represents 80% chance of detecting an

medium sized effect that genuinely existed within the data, a total sample size

of 106 participants would have been required. Gathering this many participants

was unfortunately not possible within the current study.

It was therefore decided, given the low statistical power and small sample size

in the research, which lower the chance of obtaining a Type I error, that a

Bonferroni correction would not be utilised in the following analyses.

4.1.2 Attrition

There was considerable attrition from the initial sample over the course of the

intervention sessions. Table 4.2 displays the numbers of participants who

completed measures at pre-test and post-test in each setting.

Number of participants at

pre-test

Number of participants at

post-test

School A 12 5

School B 7 0

School C 4 3

School D 6 6

School E 6 4

School F 6 5

Table 4.2 Table to Show the Number of Participants Present at Pre-test and Post-test

At post-test 56% of the original sample remained, 83% of the original control

group and only 35% of the intervention group. In School A 4 students were

removed by the school due to disruptive and bullying behaviour, 2 chose not to

attend the sessions and a further pupil left the school. School B decided not to

continue with the programme after four weeks. In School C one pupil self-

selected out of the sessions as she did not feel they were relevant to her needs

and in Schools E and F the data for 2 participants was removed when it came to

99

light that they had been accessing support for their behaviour from other

sources, which may have impacted upon the reliability of the results gathered in

this research. A further pupil in School E moved schools.

The stringent SRM-SF scoring protocols meant that some of the participant

measures were unscorable, in the intervention group, only 75% of the remaining

participants could be included in the analyses and in the control group 40% of

the participants produced scorable SRM-SF surveys.

In order to ensure that there were no differences between those who remained

in the study and those who dropped out, which may suggest a form of bias was

present in the group that remained, Mann Whitney U tests were conducted on 5

of the dependent variables from the self-report measures. This included the

social skill and problem behaviour composite scores and moral reasoning

maturity scores. These analyses found no significant differences between the

groups which suggests that those who remained in the study were comparable

to those who left the programme.

For those participants who remained in the research the response rate for

parents was 87% at pre-test and 74% at post-test. One parent did not consent

to participating in the research themselves and a further four did not access the

support which was offered by school for parents with English as an Additional

Language. Finally one parent had little direct contact with the school whilst her

pupil was excluded, which occurred during the post-measure data collection.

The response rate for teachers was 100% at pre-test and 83% at post-test. One

teacher in School D was absent from work due to long term illness at post-

measure. The implications of this attrition on the results will be discussed further

in Chapter 5.

4.1.3 Descriptive Data

4.1.3.1 Measures of Central Tendency and Variability

Tables 4.3-4.8 display the Mean, Median, Range and Standard Deviation (SD)

for each of the dependent variables gathered from each source; parent, teacher

and self-report, at both data collection points (pre and post-intervention) for the

control and intervention groups. Both the mean and median have been provided

100

as, whilst the mean is considered to be the most popular measure, the median

is less sensitive to extreme scores (Dancey & Reidy, 2011).

101

Dependent

Variable

Intervention Group Control Group

Mean Median Range SD Mean Median Range SD

Communication 8.75 10 10 3.77 10.27 12 14 4.698

Cooperation 7.5 7 12 4.309 10.6 11 17 5.18

Assertion 10 10.5 14 4.408 10.2 12 14 4.263

Responsibility 9.25 9 13 3.77 10.87 10 16 4.033

Empathy 9 10 12 3.625 9 9 15 4.375

Engagement 11.25 12.5 15 5.445 12.467 12 12 2.722

Self Control 4.63 4 10 3.204 6.33 6 15 5.327

Social Skills

Composite-

Raw Score

60.38 65 59 23.366 69.73 71 87 25.073

Social Skills

Composite-

Standardised

Score

73.25 76.5 43 16.833 80.27 82 68 19.44

Externalising 21.88 20.5 22 7.918 16.87 17 30 9.211

Bullying 7 8 11 3.742 6.27 7 14 4.044

Hyperactivity 12.63 11.5 15 4.838 10.47 9 19 6.523

Internalising 9.25 8.5 19 6.274 4.67 5 9 3.086

Problem

Behaviours

Composite-

Raw Score

40.13 41 44 16.075 30.73 30 53 16.355

Problem

Behaviours

Composite-

Standardised

Score

121.5 122 46 16.733 111.53 111 55 17.016

SRMS 200.33 194 97 34.662 240.5 219 99 40.218

Table 4.3: Table to Show the Descriptive Data for the Self-report Measures at Pre-test

102

Dependent

Variable

Intervention Group Control Group

Mean Median Range SD Mean Median Range SD

Communication 8 7.5 13 4.071 8.8 8 15 4.57

Cooperation 6.5 7 11 3.703 10.47 9 16 5.263

Assertion 9.38 8 13 3.998 11.07 12 16 4.891

Responsibility 8.63 9.5 14 4.984 12 13 19 5.141

Empathy 10.13 10.5 9 3.227 8.93 9 13 4.044

Engagement 12 13.5 12 4.598 13.33 13 14 3.735

Self Control 5.5 5 11 3.464 7.13 5 16 5.041

Social Skills

Composite-

Raw Score

60.13 60 76 24.263 71.73 71 102 28.632

Social Skills

Composite-

Standardised

Score

67 64.5 40 13.99 81.73 79 77 31.526

Externalising 22.25 26 22 8.664 13.93 16 32 9.051

Bullying 6.38 7.5 9 3.889 4.4 4 11 3.18

Hyperactivity 14.38 15 16 5.731 9.4 9 21 6.412

Internalising 8 6.5 19 6.234 5.2 4 14 4.362

Problem

Behaviours

Composite-

Raw Score

41.38 48 51 17.695 26.87 28 60 17.108

Problem

Behaviours

Composite-

Standardised

Score

122.88 130 53 18.357 107.4 109 62 17.904

SRMS 226.17 219 77 28.548 233.33 242 119 46.237

Table 4.4: Table to Show the Descriptive Data for the Self-Report Measures at Post-test

103

Dependent

Variable

Intervention Group Control Group

Mean Median Range SD Mean Median Range SD

Communication 12.63 13.5 9 3.021 16.45 19 13 4.906

Cooperation 9 9.5 7 2.828 12.27 14 12 4.735

Assertion 13.38 14 5 1.923 14 15 11 3.376

Responsibility 9.38 9 8 2.774 12.27 12 12 4.496

Empathy 8.25 8 9 3.151 10.27 8 10 4.052

Engagement 11.63 12 12 3.889 15.36 18 13 4.433

Self Control 7.63 8 12 4.502 11.73 8 18 6.944

Social Skills

Composite-

Raw score

71.88 70.5 54 17.956 92.36 103 75 28.696

Social Skills

Composite-

Standardised

score

83 82 35 11.94 96.36 104 50 19.33

Externalising 17.63 18 12 4.241 9.91 13 22 8.538

Bullying 5.38 5 5 1.768 3.36 1 10 3.931

Hyperactivity 10.88 11.5 9 3.044 7.45 11 14 5.82

Internalising 7.88 7.5 11 3.441 5.27 7 13 4.962

Autistic

Spectrum 18.63 18.5 16 5.37 11.55 15 24 8.96

Problem

Behaviours

Composite-

Raw Scores

39 40 28 9.502 23.64 31 51 20.446

Problem

Behaviours

Composite-

Standardised

Scores

134.75 134.5 30 10.938 118.27 128 67 27.236

Table 4.5: Table to Show the Descriptive Data for the Teacher Report Measures at Pre-test

104

Dependent

Variable

Intervention Group Control Group

Mean Median Range SD Mean Median Mode SD

Communication 11.5 11 8 2.928 14.18 13 13 4.6

Cooperation 7.88 7.5 9 3.227 10.64 10 13 5.537

Assertion 13.38 13.5 8 3.159 13.45 13 11 3.045

Responsibility 8 8 7 2.619 11.36 11 12 4.632

Empathy 7.38 7 5 1.768 10.09 11 14 4.826

Engagement 11.25 10.5 9 2.765 14.09 14 15 4.392

Self Control 7.25 6.5 13 4.559 10.09 8 18 5.856

Social Skills

Composite-

Raw score

66.88 65 46 16.771 83.91 79 87 30.576

Social Skills

Composite-

Standardised

score

79.63 78.5 31 11.338 90.55 87 58 20.656

Externalising 16.5 16 11 4.071 12.45 13 24 8.711

Bullying 5.75 6 3 1.035 4.55 5 10 3.882

Hyperactivity 11.88 12 5 1.885 8.09 9 14 5.504

Internalising 7.13 8 10 3.482 3.55 3 9 2.841

Autistic

Spectrum 19.25 20.5 10 4.432 11.82 11 24 8.122

Problem

Behaviours

Composite-

Raw Scores

39 41 20 7.27 24.73 29 48 17.147

Problem

Behaviours

Composite-

Standardised

Scores

134.75 135 24 8.276 119.55 126 63 22.866

Table 4.6: Table to Show the Descriptive Data for the Teacher Report Measures at Post-test

105

Dependent

Variable

Intervention Group Control Group

Mean Median Range SD Mean Median Range SD

Communication 13.5 15.5 10 3.987 15.64 16 9 2.803

Cooperation 11.17 10 5 2.229 12.27 12 9 2.687

Assertion 15 16 9 3.521 14.64 14 10 3.139

Responsibility 11.67 12 3 1.033 12.73 13 11 3.58

Empathy 11.17 12.5 14 4.916 11.64 11 11 3.414

Engagement 13 13 9 3.633 15.27 16 7 2.328

Self Control 6 5.5 11 3.742 9 8 14 3.899

Social Skills

Composite-

Raw Score

81.5 84 57 18.992 91 91 43 14.064

Social Skills

Composite-

Standardised

Score

85.33 87.5 44 14.814 93 94 36 11.287

Externalising 13 11.5 19 7.127 10.09 10 19 6.188

Bullying 3.33 2 12 4.367 2.82 2 8 2.483

Hyperactivity 10 10 10 3.464 6.91 6 13 4.764

Internalising 8.5 8 14 5.857 4.09 3 8 2.844

Autistic

Spectrum 17.5 17.5 23 7.396 10.45 9 17 4.967

Problem

Behaviours

Composite-

Raw Scores

32.17 30 55 18.978 20.73 21 38 12.507

Problem

Behaviours

Composite-

Standardised

Scores

121 118 64 22.83 107.36 106 45 14.942

Table 4.7: Table to Show the Descriptive Data for the Parent Report Measures at Pre-test

106

Dependent

Variable

Intervention Group Control Group

Mean Median Range SD Mean Median Range SD

Communication 12.33 12.5 10 3.83 13.91 15 15 4.253

Cooperation 9.17 9 6 2.229 12.55 13 8 2.423

Assertion 14 14 4 1.414 14 14 8 2.236

Responsibility 9.67 10.5 8 3.011 11.18 13 16 4.557

Empathy 9.83 10.5 8 3.061 11.09 12 10 3.239

Engagement 12.33 13 13 4.502 14.45 14 11 3.532

Self Control 4.83 3.5 11 4.07 9.82 10 14 3.97

Social Skills

Composite-

Raw Score

72.17 74.5 49 16.018 87 84 59 17.838

Social Skills

Composite-

Standardised

Score

77.67 80 38 12.437 89.55 87 46 13.917

Externalising 15.5 13 21 7.477 9.55 9 22 7.764

Bullying 3.33 3 9 3.445 2.82 2 8 2.786

Hyperactivity 11.33 11 15 5.317 6.64 5 15 5.104

Internalising 9.33 7.5 18 6.743 3.27 3 10 2.724

Autistic

Spectrum 18.83 16.5 29 10.572 11.27 11 20 5.918

Problem

Behaviours

Composite-

Raw Scores

37 31.5 65 22.724 19.45 17 46 14.376

Problem

Behaviours

Composite-

Standardised

Scores

123.67 120 57 19.775 105.91 101 60 17.835

Table 4.8: Table to Show the Descriptive Data for the Parent Report Measures at Post-test

107

4.1.3.2 Distribution and Variance of the Data

4.1.3.2.1 Normal Distribution

In order to ascertain which form of inferential statistical tests should be used the

distribution of the data was analysed. Histograms and Q-Q plots were created

and visually analysed, then statistical analyses were carried out in the form of z

scores for skew and kurtosis and Shapiro-Wilk analyses, in order to ascertain

the significance of the distribution. Whilst there were several methods to

consider, research has found the Shapiro-Wilk test to be the most powerful

(Razali & Wah, 2011). Appendix XIX displays the skew, kurtosis, z scores and

Shapiro-Wilk scores for each variable, at both time points, from the 3 sources of

data; self-report, parental report and teacher-report.

4.1.3.2.1.1 Summary of Findings

Of the 36 dependent variables, 19 were found to violate normal distribution

patterns, achieving z scores outside of the -1.96 to +1.96 range or Shapiro Wilk

significance scores below 0.05. Visual analyses of the histograms and Q-Q

plots also suggested that much of the data was skewed or bimodal. This was

expected due to the inclusion of participants who were considered to be target

individuals and role models. The creators of the SSIS-RS also state that there is

no evidence or theoretical rationale to suggest that social skills and problem

behaviours are normally distributed through the population (Gresham & Elliott,

2008).

4.1.3.2.2 Homogeneity of Variance

Homogeneity of variance refers to the assumption that data from the different

groups in the research will have similar variances for each outcome variable

(Field, 2009).

4.1.3.2.2.1 Summary of Findings

Non-parametric Levene’s tests were conducted to assess the homogeneity of

variance between the experimental and wait-list control conditions. Tables

displaying the results of these analyses can be found in Appendix XX. Five

variables were found to have significant variance between the groups, achieving

p-levels below 0.05, thus violating the assumption of homogeneity of variance.

108

4.1.4 Inferential Statistics

4.1.4.1 Assumptions of Parametric Statistical Tests

Parametric statistical tests make several assumptions about the data to be

analysed. For example they require that the data is normally distributed, there is

homogeneity of variance and the data is measured at least at an interval level

(Field, 2009) as well as the data having been taken from a random sample of

the population. Due to the ordinal nature of the majority of the data pertaining to

the SSIS-RS measure, the non-random sample and the high number of non-

normal distributions highlighted above, the researcher has decided to conduct

non-parametric statistical tests, which are less sensitive to such issues.

However, it is acknowledged that such tests are less powerful than parametric

analyses, which are more likely to find a difference in the population should one

exist (Dancey & Reidy, 2011).

4.1.4.2 Between-group Comparisons

4.1.4.2.1 Similarity at pre-test

As the participants were not randomly allocated to the conditions of the

research, analyses were conducted to see whether there were any pre-existing

differences between the control and experimental participants. Without such

tests any differences noted at post-test may have been attributed to the effects

of the intervention, when in fact the participants differed prior to their

participation.

A Mann-Whitney U test was performed on each dependent variable. 6

dependent variables, 2 from each of the sources reporting data, were found to

be significantly different. Table 4.9 displays the data for these significant results.

Source DV Intervention

group

median

score

Intervention

group size

(N)

Control

group

median

score

Control

group

size (N)

U Z P

Self

report

Internalising

behaviours

(SSIS)

8.5 8 5 15 30.5 -1.913 0.028

109

Sociomoral

Reflection

Maturity

Score

(SRM-SF)

194 6 219 6 5 -2.085 0.018

Teacher

report

Engagement

(SSIS) 12 8 18 11 22.5 -1.784 0.039

Externalising

(SSIS) 18 8 13 11 21 -1.908 0.029

Parent

report

Self control

(SSIS) 5.5 6 8 11 16 -1.722 0.045

Autism

Spectrum

(SSIS)

17.5 6 9 11 16 -1.716 0.045

Table 4.9: Table to Show the Significant Results of Between-group Pre-test Analyses

The descriptive data suggests that the children in the intervention group rated

themselves as possessing higher levels of internalising problem behaviours and

appeared to possess less mature moral reasoning ability compared to the

control children. The teachers rated those in the intervention group as having

lower skills of engagement and higher externalising behaviours than the control

group and finally parents rated those in the intervention group as displaying

lower levels of self-control and higher incidence of behaviours associated with

autistic-spectrum condition. Analyses of covariance were considered to take the

pre-existing differences into account. However, the small number of incidents

and the non-normal distribution of the data meant that this was not carried out.

These differences will be taken into account when considering the following

post-test analyses.

4.1.4.2.2 Similarity at post-test

Mann-Whitney U tests were performed on post-test data to investigate any

possible differences between the intervention and control groups following the

implementation of the programme. Fourteen variables were identified as being

significantly different, these are displayed in Table 4.10.

110

Source DV Intervention

group median

score

Intervention

group size

(N)

Control

group

median

score

Control

group

size (N)

U Z P

Self report Externalising

behaviour

(SSIS)

26 8 16 15 31 -1.876 0.031

Hyperactivity

(SSIS) 15 8 9 15 33 -1.746 0.042

Problem

Behaviour

Raw score

(SSIS)

48 8 28 15 30.5 -1.906 0.029

Problem

Behaviour

standardised

score (SSIS)

130 8 109 15 30.5 -1.906 0.029

Teacher

report

Engagement

(SSIS) 10.5 8 14 11 23 -1.747 0.043

Internalising

(SSIS) 8 8 3 11 18.5 -2.117 0.017

Autism

Spectrum

(SSIS)

20.5 8 11 11 18 -2.160 0.015

Parent

report

Cooperation

(SSIS) 9 6 13 11 10 -2.335 0.01

Self Control

(SSIS) 3.5 6 10 11 13 -2.026 0.022

Social Skills

Raw Score

(SSIS)

74.5 6 84 11 16 -1.710 0.048

Social Skills

Standardised

Score (SSIS)

80 6 87 11 16 -1.711 0.046

Internalising

(SSIS) 7.5 6 3 11 9 -2.455 0.006

Problem

Behaviours

Raw Score

(SSIS)

31.5 6 17 11 15 -1.812 0.037

111

Problem

Behaviours

Standardised

Score (SSIS)

120 6 101 11 13.5 -1.961 0.026

Table 4.10: Table to Show the Significant Results of Between-group Post-test Analyses.

Visual analyses of the descriptive data indicate that the intervention group

consistently received significantly less favourable results when compared to the

control group. The control group received significantly higher scores for social

skill scales and lower scores for problem behaviours. This pattern was present

in all 14 significant differences at post-test.

Regarding the pre-test analyses conducted previously, teacher-rated

engagement and parental rated self-control remained significantly lower for the

intervention group when compared to the control group. However, four of the six

significant differences were no longer apparent in the post-test data. Descriptive

data indicate that teacher reported externalising behaviours improved at post-

test for the intervention group whilst the control group remained the same as at

pre-test. Median scores for the parent reported Autism-Spectrum scales

suggest that the intervention group improved pre to post-test whilst the control

group deteriorated. Regarding self-reported levels of internalising behaviours,

both the control and intervention groups improved from pre to post-test, to the

point where there was no significant difference found between them by the

Mann-Whitney U tests. Similarly, in relation to the self-report measure of moral

reasoning maturity, both groups improved but the intervention group did so at a

greater rate, closing the gap. Figures 4.1-4.4 show how the intervention group

closed the gap with the control group on these four dependent variables.

112

Fig.4.1: Graph to show Changes in Median SRMS from Pre to Post-Test for the Control and

Intervention Groups

Fig 4.2: Graph to show Changes in Median Self-Reported Internalising Score from Pre to Post-

Test for the Control and Intervention Groups

113

Fig 4.3: Graph to Show Changes in Median Teacher-Reported Externalising Score from Pre to

Post-test for the Intervention and Control Groups

Fig 4.4: Graph to Show the Changes in Median Parent Reported Autism Spectrum Scores from

Pre to Post-Test for the Intervention and Control Groups.

114

4.1.4.3 Within-group Comparisons

Within-groups analyses were then completed using Wilcoxon Signed Ranks

tests in order to find out if there were any changes over the course of the

intervention. Teacher-report data revealed no significant differences from pre to

post-test for both the control and intervention groups. However, parent and self-

report measures showed five significant changes over time for the intervention

group and self-report measures revealed three significant changes for the

control group at post-test. These results are displayed in Table 4.11.

Condition Source DV Pre-test

Median

Post-

test

Median

N T Z P

Intervention

group

Self

report

Sociomoral

Reflection

Maturity Score

(SRM-SF)

194 219 6 0 -2.201 0.016

Parent

report

Cooperation

(SSIS) 10 9 6 0 -2.032 0.031

Social Skills

Raw Score

(SSIS)

84 74.5 6 0 -2.201 0.016

Social Skills

Standardised

Score (SSIS)

87.5 80 6 0 -2.207 0.016

Externalising

(SSIS) 11.5 13 6 0 -2.214 0.016

Control

group

Self

report

Communication

(SSIS) 12 8 15 16.5 -2.039 0.021

Externalising

(SSIS) 17 16 15 16.5 -2.029 0.021

Bullying (SSIS) 7 4 15 26.5 -1.912 0.029

Table 4.11: Table to Show the Significant Changes from Pre to Post-Test from Within-Groups

Analyses

These results show that the intervention group improved significantly in the

maturity of their moral reasoning responses from pre to post-test. However, the

other significant results for this group, gathered from parent-report measures,

suggest that the cooperation scale, social skills composite score, social skills

115

standardised score and externalising behaviours all deteriorated over the

course of the intervention.

The three significant differences noted in the control group indicated that whilst

they displayed lower levels of externalising and bullying behaviours, which

suggests improvement over time, the scores received on the communication

scale also decreased over time, which indicates a deterioration in this social

skill.

4.1.4.3.1 Effect sizes

Where a significant improvement was noted in the intervention group effect

sizes were calculated in the form of Cohen’s r, as this is the calculation

recommended when analysing results from Wilcoxon non-parametric tests

(Fritz, Morris & Richler, 2012). Effect sizes allow for comparison of results

between different research studies utilising different measures (Field, 2009) and

give an indication of the finding’s practical importance (Fritz, Morris & Richler,

2012).

The SRMS was the only dependent variable found to show significant

improvement in the intervention group from pre to post-measure. Cohen’s r

calculations found this effect to be large in size (r= -0.64).

4.1.4.4 Analyses of subgroups

Analyses were conducted to identify possible patterns between subgroups of

participants.

4.1.4.4.1 Outcomes for Role Models and Target Pupils

In order to investigate the possibility that the intervention may have been

differentially effective for participants identified as ‘target pupils’ and ‘positive

role models’ the pre and post data for these groups was analysed separately

using Wilcoxon Signed Ranks tests. No significant results were found for either

category, which could be due to the small number of individuals in each group,

as small samples increase the risk of a Type II error. Only 4 role models and 4

target pupils remained in the intervention group.

116

4.1.4.4.2 Outcomes for Older and Younger Pupils

The SSIS-RS offers two different versions for children of different ages, 8-12

and 13-18 years. Participants were separated using these age brackets in order

to investigate any possible patterns based on age. Wilcoxon Signed Rank tests

were conducted on the pre and post-data for the two groups using five of the

dependent variables which were considered to provide a comprehensive

summary of the complete data set; social skills composite scores, social skills

standardised scores, problem behaviours composite scores, problem

behaviours standardised scores and the SRMS. No significant differences were

found from pre to post-test for either the older or younger groups. Again it could

be that the small sample sizes were responsible for the lack of significant

findings, as there were only three ‘younger’ participants and five ‘older’

participants, which reduced to two and four respectively when considering the

parent-report data.

Descriptive data was accessed and analysed visually. Median data from both

self-report measures and teacher-report measures for the social skills

composite scores did suggest quite a large difference between the changes

from pre to post-test of the younger and older participants. Whilst the younger

participants increased their social skills composite score from a median of 68 at

pre-test to 72 at post-test the older participants scores decreased from a

median of 62 to 52 at post-test. These scores were reflected in the standardised

scores. Similarly the teacher-report measures for the same composite

suggested that the younger participants increased their median scores from 69

to 72 whilst the older participants decreased from 72 to 58. However, it is

important to remember that the researcher is only highlighting trends, these

changes were not found to be statistically significant.

4.1.4.4.3 Attendance and Performance

Spearman’s rho correlations were completed to identify any relationships that

may be present between participant’s attendance at the sessions and their

gains in the dependent variables. It was predicted that those who attended the

sessions more frequently would receive greater gains in social skills and moral

reasoning maturity and decreased levels of problem behaviours. Change scores

were created in order to compare changes in dependent variable scores from

117

pre to post-test with percentage attendance at the sessions. Two dependent

variables achieved a significant relationship; teacher-reported cooperation

scores on the SSIS-RS (r=0.765, p=0.013) and parent reported autism

spectrum scores on the SSIS-RS (r=0.754, p=0.042). Only the first of these two,

cooperation, was in the socially desirable direction expected. Figures 4.5 and

4.6 display the correlations found.

Fig 4.5: Graph to Show the Positive Relationship Between Teacher-Reported Cooperation

Change Scores and Percentage Attendance at Sessions

118

Fig 4.6: Graph to Show the Positive Relationship Between Parent-Reported Autism Spectrum

Change Scores and Percentage Attendance at Sessions

4.1.5 Summary of Findings from Statistical Analyses

In summary, testing between groups found six dependent variables, pertaining

to the measures of social skills, problem behaviours and moral reasoning, which

were significantly different between the control and intervention groups at pre-

test. Descriptive statistics suggested that all of these outcomes were less

favourable for the intervention group. At post-test 14 outcome variables from the

social skills and problem behaviours domains were significantly different

between the groups, again all in favour of the control group. However, four of

the differences observed at pre-test were no longer present in the data, which

descriptive statistics appeared to suggest were due to the intervention group

‘closing the gap’ with the control group on the scales of externalising problem

behaviours, internalising problem behaviours, autism-spectrum behaviours and

SRMS.

Within groups tests were then completed in order to assess if the differences

noted previously had improved significantly over the course of the intervention.

119

Only self-report and parent-report measures found significant results. The

intervention group’s scores on four parent-rated dependent variables, social

skills composite score, standardised score, externalising problem behaviours

and cooperation, appeared to deteriorate. However, the improvement in

sociomoral reflection maturity noted previously achieved significance, with a

large effect size of r=-0.64. Mixed results were achieved by the control group

from their self-report measures, with significant improvements in externalising

problem behaviours and bullying from pre to post-test but deterioration in scores

for communication.

Finally analyses for several independent variables within the intervention group

were completed, in order to investigate any possible effects of age, attendance

at the sessions and ‘position’ within the group (role model or target individual)

on outcomes. Very few results were statistically significant, which the

researcher attributes to the small group sizes, leading to increased risk of Type

II error. Trends in the descriptive data appeared to suggest that the younger

participants benefitted more than the older group members in relation to gains

in social skills. Attendance at the sessions was also found to relate positively

with teacher-rated cooperation, indicating improvement, and the parent rated

autism spectrum scale, suggesting deterioration in these behaviours with

greater exposure to the intervention.

4.2 Qualitative Data Analysis

4.2.1 Qualitative Data Collection Procedures

4.2.1.1 Subsidiary Research Question

Key stakeholders in the research, specifically the training provider from iCART

and the EPS, were interested in participant’s experiences of this initial pilot of

the ART intervention in order to inform future implementation. Therefore an

additional subsidiary question was explored as part of the current research:

What are the views of those involved in the initial pilot of the ART intervention

sessions, in relation to programme implementation, contents and effectiveness?

It was decided that gathering information from both group members and

facilitators would be appropriate in answering this question, as it was predicted

120

that these two groups would have differing experiences of the programme.

Therefore thematic analyses were carried out separately for these two groups.

Questions on the facilitator questionnaires included (please see copy in

Appendix XV):

Do you think the ART intervention was effective?

What, in your opinion, contributed to these outcomes? (What worked well

or did not work so well?)

If you were to run this programme again what changes would you make?

The interview with the group members was semi-structured in form and the

following prompts were used to gather information relevant to the research

question (please see Appendix XIV for a full copy):

Effectiveness; Do you think the ART group helped you? How? Have you

used what you learnt? Which part have you used the most? What parts

do you think were effective? Were there things that didn’t work for you?

Implementation; What did you think of the ART sessions? How did you

feel taking part in the groups? What did you like about the sessions/not

like? What could be changed to make ART better?

4.2.1.2 Selection of participants

All group members and facilitators who took part in the intervention sessions

were invited to take part in this additional element of the research, even if they

had not participated in the quantitative data collection procedures. Parental

consent was sent home to the parents of the 10 remaining group members and

8 were returned to the researcher. These young people were then asked to

provide their own written consent, all agreed to take part (please see

Appendices XVI and XVII for copies of these forms). Each interview was

conducted in a one-to-one setting with the researcher following the completion

of the intervention sessions and lasted between 5-10 minutes.

The six facilitators had previously signed consent forms, prior to the intervention

sessions. These forms asked for their participation in the 4-item questionnaire,

121

which was distributed via email during the week following the completion of the

intervention sessions.

4.2.2 Thematic Analysis Process

This section provides a brief summary of the thematic analysis process

undertaken, in accordance with procedures provided by Braun and Clarke

(2006).

4.2.2.1 Familiarisation with the Data Set

This initial stage required the researcher to immerse themselves in the data.

Firstly data was transcribed, either from the original audio recordings of the

interviews or questionnaires which were sent via email. These documents were

then read repeatedly in order to search for patterns and meanings which would

inform initial codes.

4.2.2.2 Generation of Initial Codes

Codes, also referred to as ‘the most basic segment, or element, of the raw data

or information that can be assessed in a meaningful way regarding the

phenomenon’ (Boyatzis, 1998, p.63), were created where semantic features of

the data appeared interesting and relevant to the research question. These

codes were written alongside each excerpt on printouts of the entire data set.

Where an excerpt fitted into more than one code multiple codes were written.

For example the excerpt: ‘I felt happy cos like it’s fun...and the people in my

group were nice...’ was coded with ‘enjoyment of sessions’ and ‘social

advantages’.

4.2.2.3 Searching for Themes

At this stage of analysis, codes were considered in comparison to one another

in order to start combining them to create themes. The cut-out extracts were

placed together where similar codes had been used and repeatedly read to

analyse any possible relationships between their contents. Where patterns

emerged, candidate titles for these themes and subthemes were written onto

post-its and placed alongside the coded excerpts (please see Appendix XXI and

XII for a picture of the original thematic networks). Some of the previously

identified codes went on to become main themes and subthemes depending on

their ability to accurately describe the excerpts contained within them. For

122

example the code ‘challenging behaviour’ went on to become a subtheme of the

theme ‘nurturing a positive environment’.

4.2.2.4 Reviewing Themes

During this phase, candidate themes were appraised in relation to the patterns

of codes and excerpts contained within and by comparing themes to one

another. Braun and Clarke (2006) state that ‘Data within themes should cohere

together meaningfully, while there should be clear and identifiable distinctions

between themes.’ (p.20). For example, the subtheme ‘developing positive

relationships’ was previously associated with the theme ‘skills/strategies

employed by the facilitators’. However, as some of the excerpts described the

relationships between group members, not facilitators, it was decided that this

subtheme was actually more applicable to the theme ‘nurturing a positive

environment’. Some subthemes also became themes themselves, for example

‘selection of participants’ was previously a subtheme of ‘group composition’ but

because it appeared to contain subthemes of its own it was promoted to a

theme.

As themes should represent the most ‘...salient constellations of meanings

present in the data set.’ (Joffe, 2012, p.209) themes where there was

insufficient data to support them, for example where they were only represented

by a single utterance from a single participant, were removed from the analysis.

Whilst this means that potentially interesting and relevant information may have

been omitted it ensures that the findings remain coherent and concise. Where

participants had responded with a short closed statement which provided little

information for analysis, for example stating ‘yes’ or ‘no’ in response to the

question ‘Do you think the ART group helped you?’, the researcher decided that

these would be counted and reported separately to the themes, in order to

ascertain the relative weight of the statements in relation to the number of

participants who held the same view.

Following these refinements, the validity of the individual themes was then

appraised in relation to the entire data set. The researcher re-read the original

transcripts and ensured that the themes accurately represented the meanings

present when the data set was considered as a whole.

123

A thematic map was created for each data set (Figures 4.7-4.10) in order to

provide a visual conceptualisation of the patterns of data.

4.2.2.5 Naming and Defining Themes

Finally themes were operationalised and provided with a descriptive label in

order to identify ‘...the essence of what each theme is about.’ (Braun & Clarke,

2006).

4.2.2.5.1 Inter-rater Agreement of Themes

In accordance with procedures proffered by Joffe (2012) 20% of the total data

set was analysed by a second rater, a doctoral research student with previous

experience in qualitative methodology. These procedures involved the second

rater identifying which theme or subtheme each extract belonged to. Inter-rater

agreement was 100% which is above the recommended level of 75% (Joffe,

2012). This suggests that the themes are a reliable representation of the

excerpts within.

4.2.3 Themes Constructed from the Group Member Interview Data

2 overarching themes, containing 12 themes and 9 subthemes, were identified

within the interview data from group members. Figures 4.7 and 4.8 display

these themes as thematic maps. In the following sections themes will be

outlined in detail, organised by the overarching themes.

124

Fig 4.7: Thematic map of the overarching theme ‘Reported Outcomes Experienced’ and the associated themes generated from group member interview data in response to the question: What are

the views and experiences of those involved in the initial pilot of the ART intervention sessions, in relation to programme implementation, contents and effectiveness?

125

Fig 4.8: Thematic map of the overarching theme ‘Perceptions of Intervention Contents and Implementation’ and the associated themes generated from group member interview data in response to

the question: What are the views and experiences of those involved in the initial pilot of the ART intervention sessions, in relation to programme implementation, contents and effectiveness?

126

4.2.3.1 ‘Reported Outcomes Experienced’

This overarching theme describes instances where interviewees described

changes that had occurred which they attributed to their participation in the

intervention. In relation to the data which was counted as opposed to analysed

thematically, when participants were asked if the ART groups had helped them

all 8 responded affirmatively. Four themes were devised which together

describe the outcomes experienced by the participants following their

involvement in ART.

Management of angry thoughts

Participants described ways in which the intervention had helped them to

manage angry thoughts. This theme included two sub themes:

Calming down

Acquisition of explicitly taught anger management techniques

Several participants made reference to the intervention helping them to ‘calm

down’. Some expanded upon this, suggesting that it was this calmness that was

helping them to avoid behaving in a manner which could be considered

aggressive. For example Participant A stated: ‘That if someone like says

something about me I just calm down and never go and fight them’. The

majority of the participants also made reference to their application of the anger

management techniques which were taught during the intervention sessions.

Participant F commented: ‘I think twice, is this the right answer I should be

giving to the teacher or to this pupil and what’s gonna be my consequence?’

and Participant C mentioned the self-talk process: ‘...I used the thinking well

bubble process and then like I just didn’t get angry...’

Increased social skills

Participants also described the social skills which they felt they had gained and

used since the intervention, for example Participant H stated: ‘In one lesson

we’ve been like learning like to give compliments and I’ve been giving

compliments to like my friends...’

127

Positive responses of others outside the sessions

Participant H continued, commenting that since this change in his behaviour his

friends ‘...seem more happier as well’. According to Participant H others had

also noticed the changes in his behaviour including ‘...my family is more proud

of me because I’m not getting detentions as often’.

Additional personal benefits

In addition to positive outcomes directly associated with the primary goals of the

intervention, participants highlighted several personal benefits which they

attributed to their participation in ART. These were placed under three

subheadings:

Social and developmental benefits of being part of a group

Feeling more mature

Reduced sanctions/increased merits

Several participants mentioned that there had been social benefits to

participating in a group. Participant B commented that the group had provided

him with friends from other year groups who he previously hadn’t spoken to: ‘we

built like a friendship group so like in other years I know who they are and stuff.’

The ‘developmental’ aspect referred to Participant A’s comment that:

‘sometimes we messed about...but...when we had actually did learn we done

something we like we would come up with good things’, suggesting that as a

group, when they worked together there were benefits to their learning and

development. Participant A also noticed further positive changes in himself:

‘...I’m ...starting to like get a little bit, like, more mature now.’ Finally Participants

A and H noted that they were receiving less sanctions at school, with Participant

A stating: ‘...cos before I used to get like detentions, detentions but I’m not as

much now.’

4.2.3.2 ‘Perceptions of Intervention Contents and Implementation’

The second overarching theme provides an insight into the group member’s

views about the contents of the ART sessions and elements of implementation

128

which contributed to the intervention’s effectiveness and their enjoyment of the

sessions.

Success of the Moral Reasoning component

Half of the participants identified the moral reasoning sessions as either the

most effective or most enjoyable part of the intervention programme. Participant

B stated: ‘...cos moral reasoning’s like debating a bit so I like doing that....’.

ACT sessions were the least popular

In contrast the ACT component was viewed as the least successful. Participant

C described an occasion where an activity from these sessions, which involved

sharing experiences where they felt angry, ‘...didn’t work that well...’ because ‘It

just caused like just over-talking and then everyone started like saying stuff they

didn’t like and then [the facilitators] couldn’t obviously talk.’

Sessions were enjoyable experiences

When asked how they felt taking part, half of the participants were explicit about

their enjoyment of the sessions, for example Participant D commented: ‘I felt

happy cos its like fun’ and Participant H stated: ‘I enjoyed it and um it did help a

lot and erm I’m happy for that...’.

Popularity of the games

Two participants mentioned the games when asked if there were any particular

activities which they enjoyed during the sessions.

Difficulties caused by group size

Contradicting views were gathered under this theme. Whilst one participant

highlighted the need for more group members, another mentioned negative

outcomes which they associated with taking part in an initially larger group.

These views were captured using two subthemes:

Larger groups would negate the difficulties caused by non-attendance

Larger groups lead to increased problem behaviour

129

The first stemmed from Participant E’s comment ‘...there’s only a few of us and

then sometimes take days off...we need like more people who want to

come...cos there’s only a few of us so we can’t contribute that much’. In

contrast, Participant F recalled the difficulties he had experienced when his

group had been larger: ‘...it was a bigger group....there was more shouting out

and you couldn’t really um understand the questions...I’d try having a smaller

group, about six or so on.’

Contents was relevant to the real world

Participants felt that the activities, particularly role plays and moral reasoning

discussions, gave them opportunities to practise their skills and made them

think about things that might occur in the real world. Participant F also

appreciated the opportunity to problem solve ‘bad things’ that had happened to

them that day so that in future ‘...we could er stop them from happening again.’

Mixed feelings about missing lessons

A second theme which included some contradictory views was that of missing

lessons:

Missing lessons is good

Missing favourable lessons initially leads to regret

Whilst Participant A stated that this made him feel ‘good’ about coming to

sessions every week, Participant F stated that missing his favourite lessons was

a disadvantage of the sessions which initially meant he ‘...regretted coming..’.

However, over time he began to value the sessions, stating: ‘...looking back uh I

don’t regret coming to them it’s helped me.’

Enjoyment of role plays which provided opportunities to practise new skills

In addition to the games, three participants also highlighted the role plays as

being an activity which they enjoyed, with Participant H commenting that it

illustrated to him ‘...how we could do something better’ and Participant G

suggesting that ‘...more drama...’ would be a beneficial future improvement.

130

4.2.4 Themes Constructed from the Facilitator Questionnaire Data

2 overarching themes, encompassing 9 themes and 27 subthemes, were

identified within the questionnaire data provided by the group facilitators.

Figures 4.9 and 4.10 display these themes as Thematic Maps.

Fig 4.9: Thematic map of one overarching theme and the associated themes generated from

facilitator questionnaire data in response to the question: What are the views and experiences

of those involved in the initial pilot of the ART intervention sessions, in relation to programme

implementation, contents and effectiveness?

131

Fig 4.10: Thematic Map of the Second Overarching Theme and Associated Themes Generated from Facilitator Questionnaire Data in Response to the Question: What are the views

and experiences of those involved in the initial pilot of the ART intervention sessions, in relation to programme implementation, contents and effectiveness?

132

4.2.4.1’Reported Outcomes’

This overarching theme encapsulated several themes which describe outcomes

which the facilitators attributed to the intervention.

Evidence of positive changes internally

A theme was created which contained all of the references to evidence which

suggested pupil’s had changed in their internal state. Two subthemes

represented different types of change; increased knowledge and development

of more positive attitude and greater ability to reflect upon their difficulties:

Positive changes in attitude/greater reflection

Knowledge of taught skills, strategies and information

Illustrating the first subtheme Facilitator B stated: ‘...changes in attitude during

group discussions were noted over the programme and within the group

students did reflect well on their own social and moral issues’. In regard to the

second subtheme facilitators referred to participant’s knowledge of explicitly

taught information for example: ‘pupils were able to demonstrate knowledge of

the anger circle’ (Facilitator F) and specific taught skills ‘[pupil’s name]

appeared to develop more mature moral reasoning’ (Facilitator A).

Benefits for the young person themselves

Similar to the group member data, the facilitators also noted some additional

personal benefits which they attributed to the pupil’s participation in the

intervention. These were grouped into two subthemes:

More positive view of their future

Positive responses of others/increased opportunities

Facilitator A commented that one pupil in her group ‘...now sees himself as

having a positive future and he could now allow himself to be more aspirational’.

It was also mentioned that this young person was receiving greater

opportunities outside of the sessions due to his changing behaviour: ‘teachers

133

being kinder to him, school offering him lots of opportunities to get involved in

extra-curricular activities.’

Observable positive behaviour change

This theme referred to any instance where a facilitator made reference to

observable behaviour change following the intervention. Four subthemes were

developed:

Increased use of pro-social skills

Positive influence on others

Management of anger/lowered aggressive behaviours

Positive changes noticed by others outside of the sessions

During the sessions facilitators noticed changes in the participant’s social skills

for example Facilitator D wrote: ‘A few pupils started the programme saying

very little and avoiding eye contact. Their level of engagement was limited. The

same pupils became articulate in sessions, fully engaging and had a positive

influence on the cohesion in the group’. This influence on others was also

mentioned by Facilitator A, who commented: ‘[participant name] became a role

model by the end and helped encourage the right behaviours from within the

group’. The intervention was viewed as effective in supporting pupils with

management of anger and aggression, for example Facilitator A wrote: ‘ART

was effective in enabling [name] to better manage his aggression. This was a

unanimous view’. Several of the facilitators referred to the information they had

received from others who had observed the young people in the school

environment, for example Facilitator C stated: ‘reports from school staff were

positive for the young people who completed it’ and Facilitator A commented:

‘the changes in his behaviour have also inspired other pupils to ask to

participate in ART.’

4.2.4.2 ‘Factors Impacting Upon Success’

In response to the question: ‘Do you think the ART intervention was effective?’

all 6 facilitators confirmed that the intervention was partially effective, in that it

helped some of the students or helped the students with some of their

134

difficulties, but not all. This overarching theme highlights factors which the

facilitators believe impacted upon the success of the programme and

suggestions for improvements to inform future applications.

Skills/strategies employed by the facilitators

Several of the extracts referred to strategies which the facilitators chose to

employ and comments about the effectiveness of such strategies

Differentiation to make content realistic

Use of rewards

This second subtheme contained contradictory comments. For example whilst

Facilitator F wrote: ‘Rewards such as sweets for obeying the rules had limited

success’. Facilitator C stated: ‘individualised praise....helped to build their

confidence and ....increased their motivation to attend.’ This suggests that the

verbal, individualised praise may have been more effective than physical

reward. In relation to differentiation, Facilitator F wrote: ‘the trainers had to

‘toughen’ up the examples used to make them appear realistic to the young

people.’

Contents of the intervention sessions

Facilitators also provided their views on the contents of the intervention

programme, which revealed 4 subthemes:

Issues regarding accessibility of language

Importance of games to provide enjoyment and relationship building

opportunities

Role plays supported participant’s understanding and applied to real life

Moral reasoning sessions increased enjoyment and participation

A number of facilitators felt that the complex concepts and associated language

were sometimes too difficult for the participants to access, for example

Facilitators D and C respectively commented: ‘the anger control component

needs to be changed with more child friendly language’ and ‘I would adapt the

135

programme...to meet the needs of the group and to ensure they can access and

understand the contents of it’, going on to suggest that an ‘average level’ of

understanding and language was required. Similar to the young people, the

facilitators noted the success of the games, particularly in regard to relationship

building: ‘The use of games and activities that helped the group get to know

each other and us’ (Facilitator C). They also mentioned the effectiveness of the

role plays, in keeping with the group member’s views previously: ‘participants

enjoyed role play which enabled them to understand the process’ (Facilitator D).

Finally Moral Reasoning sessions were identified as having ‘worked well’ or

being ‘most successful’ as they ‘...generated wider discussions about issues

relating directly to them which were very positive discussions’ (Facilitator B).

Selection of participants

Pre-assessment of within participant characteristics needed to assess

suitability for the programme

Compulsory participation reduces ‘buy in’

Facilitators’ felt that the intervention was more beneficial for participants with

particular characteristics including: ‘...motivation to change’; ‘self-identification of

need of the programme’ and ‘ability to comply with the programme’ (Facilitator

B), some suggested that particular forms of problem behaviour, for example

‘instrumental aggression’ and ‘non-compliance’ were not as applicable as

‘aggression’ (Facilitator A) and that in future these things should be assessed in

advance to ensure the participant’s ‘suitability’ (Facilitator B). Secondly a

number of facilitators mentioned issues with the participants having been

selected or ‘forced’ to partake, which they considered to have reduced ‘buy in’

(Facilitator A). Facilitator F stated ‘the young people need to choose to be part

of the project’, Facilitator E also described a specific example of difficulty that

arose: ‘The identified pupil was aware that he had to complete course and

trainers could not exclude-he stated this publically’.

136

Organisational factors

The most salient theme, in relation to the number of extracts which contributed

to it, was that of organisational factors and the way in which the following

subthemes impacted upon success of the intervention:

Scheduling issues

Need for recognition of the programme at a whole-school level

Engagement of support outside sessions required

Participants held negative perceptions of the intervention setting

Facilitators should be provided with authority to use school’s merit

systems

Several facilitators mentioned different scheduling issues, which led to

participants not attending the initial meet and greet sessions and pupils

‘...opting out...’ because of clashes with exam subjects. Facilitator E suggested

that the timetabling had been structured around room availability and not pupil’s

timetables. In relation to recognition of the programme, suggestions for future

improvements included ensuring achievement was celebrated at a whole school

level and ‘SMT need to show a commitment to the project’ (Facilitator F).

Parents and teachers were mentioned in relation to engaging support outside of

the group. Facilitator E mentioned that there had been difficulty gaining parental

consent and Facilitator D suggested the use of introduction sessions for parents

so that skills can be practised at home. Another suggested a similar information

session for staff at the start of the programme to encourage them in ‘supporting

the young people in generalising the skills they were learning’ (Facilitator A).

Student’s perceptions of the room that was allocated was highlighted as a

barrier to effective implementation, Facilitator E recalled a student expressing ‘I

feel like a retard coming here’. Finally, whilst the use of the school’s reward

system ‘...worked well...to maintain motivation and interest.’ (Facilitator B) for

one group, another (Facilitator F) lamented the lack of authority which meant

that they could not utilise such pre-existing systems.

137

Group composition

Three subthemes encapsulated the theme ‘group composition’:

Optimal group size=6

Heterogeneity of need is important including a balance of role models

and target individuals

Role models need high status

Six participants was suggested as an appropriate number for effective

implementation. Facilitator D stated that this allowed them to ‘...focus on content

as opposed to behaviour management’. They felt that removing participants

because the group was too large was described to lead to ‘a negative dynamic

as they came to perceive us as frequently ‘kicking out’ members who did not

behave. The possibly felt under threat...’ (Facilitator C). Several facilitators also

suggested that the group should be composed of differing levels or type of

need, for example Facilitator B stated: ‘a group of non-compliant students is

likely to be ineffective’ in future a ‘balance within the group of target pupils with

social and angst issues and positive role models’ should be considered. Some

also provided suggestions as to how the role models should be selected stating

that they need to be: ‘...older...’ (Facilitator F) and ‘...high status’ (Facilitator B).

Nurturing a positive environment

This theme describes instances where facilitators referred to interpersonal

strategies which made the group environment feel positive and safe and times

when problem behaviour impacted upon the success of the sessions:

Developing positive relationships and interpersonal support

Enabling participants to feel safe, valued and listened to

Challenging negative behaviour

Firstly positive relationships were highlighted as an important factor contributing

to success, both between the facilitators and group members. Facilitator A

stated: ‘us as facilitators fostering good relationships with the young people was

138

essential.’ but also within the group: ‘they learned much from each other and

became supportive of one another by the end.’ Making the group members feel

valued was also considered important: ‘I think they valued being listened to and

not judged’ (Facilitator C). However, some did describe challenging behaviour

which they encountered that threatened the positive environment for example:

‘The behaviour of the young people was the principle difficulty. They took

advantage of the trainers trying to be approachable’ (Facilitator F). Another

trainer stated that what worked well was ‘trying to create a ‘safe’ environment

where inappropriate behaviour...is challenged and dealt with’ (Facilitator C).

4.2.5 Summary of Findings from Thematic Analysis

Qualitative feedback from those involved in this initial implementation of the

ART intervention revealed several interesting findings which can be used to

inform future application. In relation to ‘effectiveness’ a range of positive

outcomes were identified, from evidence of the taught skills being put into

practise, both in terms of changes in thought and observable behaviours, but

also positive responses of others and personal benefits for the individual.

With regard to the participant’s experiences and views of the intervention’s

content and implementation, which was the information most sought after by the

service and training provider to inform future practice, there were several salient

themes and subthemes which were evident in both the group member and

facilitator feedback. These included the success of games and role plays, the

relevance and enjoyment of the moral reasoning sessions, issues regarding the

intervention’s implementation at the organisational level including difficulties

with scheduling and finally consideration of optimal group composition.

Furthermore, in terms of the group member’s perceptions of the contents of the

sessions, themes arose which suggested that whilst the anger control sessions

were considered the least successful of the three intervention components, the

young people appreciated the strategies and teaching which were applicable to

the real world and the opportunities to talk about current scenarios which they

were finding challenging.

Other themes identified in the facilitator questionnaires suggested that an

informed selection process, additional strategies employed to tailor the sessions

139

and motivate students, as well as nurturing a positive environment were all

factors which were associated with successful implementation of the

programme and positive outcomes.

140

5. Discussion

5.1 Introduction

This chapter considers the findings derived from the current research project in

further detail and their implications for future practice and research. Firstly the

findings pertaining to each research question and the associated hypotheses

will be described, before relating these findings to theory and previous research.

A section evaluating the methodology adopted in the research will be followed

by a discussion concerning the implications of the findings for future research

and practice. Finally conclusions will be drawn summarising the outcomes of

the project and highlighting the unique contribution of the research.

This study consisted primarily of a quasi experimental design, employed to

investigate the impact of ART: A multi-component CBT-based intervention

aiming to promote social competence. Measures of problem behaviours, social

skills and moral reasoning were employed to determine the outcomes from

three sources; teachers, parents and pupils themselves. In order to provide

feedback to the stakeholders involved in this initial implementation of ART

within English secondary schools, supplementary qualitative data was gathered

to explore the perceptions of those involved in this project, specifically in regard

to factors which they felt had an influence on the success of ART.

5.2 Reflections on Quantitative Findings

5.2.1 Research Question 1

Does participation in ART, implemented in school settings in the UK,

support adolescent participants in improving their use of pro-social

behaviours and decreasing the experience of problem behaviour?

The experimental hypotheses associated with this question appeared to have

face validity given the previous research findings. The majority of the research

either reported increased levels of social skills (Coleman, Pfeiffer & Oakland,

1992; Koposov, Gundersen & Svartdal, 2014), decreased levels of problem

behaviours (Currie et al. 2009: Jones, 1991) or both (Glick & Goldstein, 1987;

Gundersen & Svartdal, 2006/2010; Langeveld, Gundersen & Svartdal, 2012).

141

This study aimed to contribute to this positive evidence base, by investigating

the impact of the ART intervention, conducted by newly trained facilitators from

the EPS, within secondary school settings in the UK.

Pre-test scores revealed differences between the intervention and control

groups on 5 subtests of social skills and problem behaviours, favouring the

control group. At post-test this increased to 14 different subscales and

composite scores, with the control group receiving more socially desirable

results consistently. Three of the pre-test differences were no longer present at

post-test and the descriptive data suggested that this was due to the

intervention group ‘closing the gap’ with the control group on problem

behaviours subscales including internalising behaviours, externalising

behaviours and autistic spectrum related behaviours. However, these pre to

post-improvements for the intervention group were not to a significant degree.

The present study found that the intervention group did not improve significantly

in either social skills or problem behaviours across the course of the

intervention. Results from parental report measures suggested that the

intervention group deteriorated in relation to the subscales of cooperation and

externalising behaviours as well as the overall social skills composite scores. In

comparison the control group, who attended their typical lessons for the

duration of the project, improved on self-reported externalising behaviours and

bullying scales. However, they also deteriorated in their self-reported use of

communication behaviours.

Additional analyses also found no support for the suggested hypotheses

regarding the differential impact of ART for subgroups based on age or whether

the group member was identified as a ‘role model’ or ‘target individual’.

Although trends in the median data did appear to suggest that the younger

participants benefitted more than the older members in relation to gains in

social skills. Attendance at the group sessions was found to correlate positively

with teacher-reported cooperation and parent reported Autism Spectrum

behaviours, which suggests that whilst increased attendance was related to

greater cooperation at school, it was also associated with higher levels of

autism-type behaviours in the home environment.

142

In this instance it would appear that the null hypothesis has been supported. It

is possible that these findings have been influenced by a number of factors.

Firstly, it is possible that the ART intervention may not be effective in improving

the social skills and problem behaviours expressed by adolescents when

implemented within UK secondary school contexts.

Secondly, issues related to the implementation process may have impacted

upon the lack of positive findings in the intervention group. For example,

anecdotal evidence suggested that participants being late to sessions and

having to fit in with the school timetable meant that activities were often cut

short. Pupil absence may also have been a barrier, with some pupils attending

as few as 53% of the sessions. Whilst the correlational analyses found that

attendance at the sessions had little association with outcomes for the

individuals themselves, it is possible that this hindered the activities which often

had to be completed in groups. As one participant stated:

‘...that’s why we need more people cos there’s only a few of us so we can’t

contribute that much.’

-‘Participant E’.

Time scale issues also meant that an originally planned follow up measure was

not possible. Previous research has found evidence of further gains 4 weeks

from the post-measure (Langeveld, Gundersen & Svartdal, 2012). The timing of

the intervention programme within the academic year also meant that the pre-

measures were taken very close to the start of term. Anecdotally the contact

personnel at the schools expressed concerns that the teachers had not had the

opportunity to get to know the students and extreme problem behaviours had

not had chance to manifest.

It is also important to note that this was an initial pilot with newly-trained

facilitators. Whilst the high treatment fidelity ratings described in section 3.5.2.3

would suggest that poor adherence to the programme did not influence the

results greatly, it may be that the common elements missed, for example

homework tasks or practising the role plays multiple times, were essential for

success. Previous research has reported that homework is significantly

143

positively related to the outcomes of therapy for anger management

(Sukhodolsky, Kassinove & Gorman, 2004). Further implementation issues and

the implications for the findings will be discussed in section 5.3.

Characteristics of the group members may also have influenced the findings.

The control group contained participants from a wider range of ages and this

may have contributed the results gathered. For example, the older students in

the control group may have been more able to reflect upon their behaviour once

they had been told they were to be part of an intervention, leading to

improvements at post-test.

It has been suggested that CBT-based programmes require more advanced

levels of cognitive development (Durlak, Fuhrman & Lampman, 1991) and the

ability to think introspectively (Beck, 1991). The high levels of intervention group

members who spoke English as an additional language and evidence from the

facilitators in the questionnaire suggesting that the language and concepts used

in the intervention programme were difficult for some of the participants to

access, may also have contributed to the lack of positive outcomes. For

example:

‘I would adapt the programme (particularly the posters) to meet the

needs of the group and to ensure they can access and understand the

content of it.’

-‘Facilitator C’

The facilitator questionnaires also revealed that poor behaviour was sometimes

a problem, leading to participant’s exclusion from the group, which may have

affected the group dynamics and hindered their ability to work successfully as a

group:

‘[having a large group] led to us having to remove group members and created

a negative dynamic as they came to perceive us as frequently ‘kicking out’

members who did not behave. They possibly felt under threat of this for at least

some of the programme.’

-‘Facilitator C’

144

Some of these areas will be discussed further in sections 5.2.3 and 5.3.

A further factor which may have contributed to the unexpected direction of

findings is that the intervention group gained increased self-awareness of their

behaviour over the course of the intervention, leading to more accurate

responses in their post-test measures. In keeping with this hypothesis one

participant commented:

‘At first um like I didn’t like know like why I was, why I needed to do it and like

um er I did like I was getting into trouble in school and in lessons and then um I

think that’s why I went like, to help me’

-‘Participant H’

The control group, having been alerted to the fact that they were to take part in

an intervention to improve their social competence, may have devised self-help

strategies or experienced increased feelings of containment as they knew they

would soon be receiving help, which lowered the perceived problem behaviours.

In keeping with this explanation, the parents and teachers may also have

become more sensitised to the behaviour of the target pupils in the intervention

group following exposure to the pre-measure. This may have led to greater

scrutiny and awareness of that child’s behaviour over the following weeks,

increasing their negative parental scores at post-test and lack of change in the

teacher measures. Studies have determined that there is often low agreement

between raters when reporting on social behaviours (Renk & Phares, 2004). For

example, in keeping with the current findings, Quinn et al. (1999) found that

those in the school context provided higher treatment ratings than parents.

Finally the measures may have had insufficient validity and reliability to

ascertain any changes. One possibility is that the intervention group member’s

responded in a socially desirable manner, attempting to create a good first

impression at pre-test and then offering more honest responses following the

intervention, once they had had greater opportunities to reflect upon their

behaviour as part of ART. Similarly, after being told that they were to be part of

a group, the control group may have responded more favourably on subsequent

145

measures to avoid having to take part in such intervention. Issues relating to the

measures will be discussed further in section 5.4.

5.2.2. Research Question 2

Does ART, implemented in school settings in the UK, contribute to the

development of adolescent participant’s moral reasoning ability?

In the present study, participants in the intervention group significantly

increased their moral reasoning maturity scores from pre to post-test whilst

control participants attending normal lessons did not. This change was found to

be large in size (r=-0.64), whereby significant differences in SRMS at pre-test

between the intervention and control groups, in which the control group was

superior, was no longer apparent at post-test.

Given the low chance of a Type I error and the large effect size, this finding

suggests that those who participated in the ART intervention experienced

positive changes in moral reasoning ability, providing support for the

experimental hypothesis. However, a definitive statement regarding causation is

not possible given competing explanations for these results in relation to

methodological issues, which are discussed further in section 5.4.

It could also be suggested that, due to the intervention group having

significantly lower SRMS scores compared to the control group at pre-test,

there was a ‘ceiling effect’ in the control group, whilst the more immature

intervention group had more to gain from the moral discussion experiences.

This explanation is in keeping with the findings of Gibbs et al (1984), whereby

participants who achieved more mature Stage 3 type moral reasoning at pre-

test did not improve over the course of moral reasoning discussion intervention,

whereas those categorised as the more immature Stage 2 reasoning at pre-test

improved significantly.

5.2.3 Consideration of the Quantitative Findings in Relation to Previous

Research

Previous research into the impact of ART has been described in Sections 2.4

and 2.6. The findings of these studies will now be compared to those of the

146

current research with the main focus upon the similar research studies identified

in the systematic literature review.

5.2.3.1 Social Skills and Problem Behaviour Outcomes

The current study found very limited evidence regarding the impact of ART on

improving social skills and problem behaviours which is in stark contrast to the

statistically significant improvements reported by the majority of the previous

research. Factors of the research which may have contributed to these

differences will now be discussed.

Firstly it would appear that there has previously been no research into the

effects of ART within a UK school context. Whilst a few studies have been

conducted in countries other than those from which the programmes originated,

for example Australia (Currie et al, 2009; Jones, 1991) and Russia (Koposov,

Gundersen & Svartdal, 2014), it could be suggested that the content of the ART

programme is culturally bound, making it less relevant to students outside of its

country of origin. Participants in the study by Currie et al (2009) commented

upon the language being specific to America, leading them to suggest that

future implementations adapt the language to suit the participant’s vernacular.

In all of the previous research the participants could be considered to have ‘high

level’ needs, whether identified by a screening measure (Gundersen & Svartdal,

2006/2010; Jones, 1991; Langeveld, Gundersen & Svartdal, 2012) or attending

youth custodial settings or institutions for pupil’s with behavioural needs (Glick &

Goldstein, 1987; Coleman, Pfeiffer & Oakland, 1992; Currie et al. 2009;

Koposov, Gundersen & Svartdal, 2014). In comparison to these groups the

current sample could be considered to have ‘lower level’ needs, which may

have impacted upon the effects of the intervention, given that previous research

has found that those with greater needs at pre-test benefit more from the

intervention (Langeveld, Gundersen & Svartdal, 2012).

However, close inspection of the average standardised scores for the social

skills and problem behaviour data do suggest that the intervention group were

perceived to be outside of the ‘average’ range on five of the six standardised

scores reported. This indicates that the measure found the young people were

experiencing elevated levels of problem behaviours and lowered social skills

147

compared to an age related sample, which suggests that initial level of need did

not contribute to the lack of significant results.

Further participant characteristics may also have contributed to the contrasting

results. Firstly the current intervention group consisted of students with a mean

age of 13 years. Whilst positive outcomes have been reported in older samples

(Currie et al 2009), studies which investigated the moderating effect of age on

intervention outcomes found that younger participants (Koposov, Gundersen &

Svardal, 2014), including those of primary school age (Langeveld, Gundersen &

Svartdal, 2012), experienced the greatest benefits from participation in ART.

This could be due to the contents of ART being more applicable to younger

students or perhaps behaviour difficulties being less ‘severe’ and more

malleable in the younger years. As Goldstein stated ‘...catch it low to prevent it

high...’ (Goldstein, 1999, p.2). These findings provide additional support for the

drive for early intervention for social, emotional and mental health difficulties

within school settings (DCSF, 2008; DfEE, 1997; DfES, 2003; Gable, Bullock &

Harader,1995).

It could be suggested that the current sample were a little old for the

programme, with one facilitator reporting that participants called the contents

‘...babyish’. Support for this notion is found in the trends from the average data,

whereby participants were divided into groups based on age. The younger

group aged 8-12 years in the current study were found to increase their median

composite social skills score from pre to post-test whereas the older participants

experienced a decrease in scores. However, it is important to note that this

trend was not statistically significant.

The current sample also had very few female participants. Following attrition

only one female remained in each condition. Whilst significant improvements

have been reported in all-male samples previously (Currie et al 2009; Glick &

Goldstein, 1987) research does suggest that gender effects are present in ART

outcomes (Langeveld, Gundersen & Svartdal, 2012). It should be noted that

these results were confounded by a floor effect, highlighting the need for further

investigation into the efficacy of ART with different genders.

148

Regarding the implementation of the intervention, diffusion of treatment effects

was a prevalent issue in previous research (Gundersen & Svartdal, 2006/2010;

Jones, 1991; Koposov, Gundersen & Svartdal 2014). As each of the five groups

who completed the current study attended different provisions, they were

unlikely to have come into contact with one another, eliminating the possibility of

treatment diffusion effects. Whilst this has negative implications for the internal

validity of the conclusions drawn (discussed further in section 5.4) it may also

be the case that changes within the peer groups outside of the ART sessions in

previous research supported them in maintaining the improvements that they

were making within the groups, something which the current cohort were

lacking.

The current implementation could almost be considered to be represent a

‘within child’ model of intervention, whereby support is offered to the young

people considered to be aggressive, as opposed to a more holistic package,

with considerable focus also being given to improving the functioning of the

systems surrounding the young person (Bronfenbrenner, 1979). A lack of

change in contextual support, including both peer and parental behaviour

(Sofronoff, Attwood & Hinton, 2005; Southam-Gerow & Kendall, 2000), make it

difficult for the young person to maintain positive behaviour changes. It may be

beneficial to intentionally include such social support within the intervention

package. For example Soloman and Wahler (1973) found that whilst peers can

reinforce the deviant behaviours displayed by others, they can also modify

these behaviours in a socially desirable direction when provided with adult

guidance.

In previous applications of ART the facilitators were predominantly members of

staff from the schools who had been trained in the programme. This research

represents the second recorded attempt to utilise facilitators who were external

to the setting in which the intervention took place. Pre-existing relationships with

the young people and prior knowledge of the supporting systems and structures

within the schools may have improved the implementation of the intervention

and, in turn, the outcomes for the young people. Organisational structures and

support will be discussed further in section 5.3.2.

149

In regard to the studies identified in the systematic literature review (section 2.6)

it was highlighted that a considerable number of the significant positive

outcomes had been gathered using measures which had not been validated,

were created by the programme developers and were closely aligned with the

programme. Studies which were considered to be of high quality only found

changes on the custom measures devised by Goldstein et al (1987), despite

employing additional validated measures. These measures have been

described as ‘...subjective and ... inaccurate...’ (Jones, 1991, p.97). Overall only

two of the five studies reported significant improvements for the treatment

groups which stemmed from independent, valid measures (Currie et al. 2009;

Gundersen & Svartdal, 2006). When these findings are considered in light of the

results of the current research it could be suggested that, whilst ART does have

a significant impact in changing the specific skills practised as part of the

intervention, the intervention does not support the young people in generalising

and developing this knowledge to support them in the execution of wider forms

of social behaviour in the short term.

Only one of the studies in the systematic literature review gathered data from

three sources (Gundersen & Svartdal, 2006). In contrast to the current

research, this previous study found significant increases in social skills and

decreases in problem behaviours on parental and teacher-report measures. It

could be suggested that the current group members failed to generalise the

skills learnt within the sessions to the wider social environments. The

intervention handbook (Gundersen, Finne & Olsen, 2006) provides guidance to

support the transfer of skills outside of the sessions. This includes the use of

homework, which was found to be employed inconsistently in the current study

and enlisting the support of those around the child, including teachers and

parents. The facilitator questionnaires suggested that this support was not

always available (please refer to section 5.3.2 for further discussion regarding

organisational support). Researchers suggest that training the skills in multiple

settings and including additional trainers, such as peers and parents, also

positively influence the use of behaviours outside of the training setting

(Cartledge & Milburn, 1995; Frederickson & Simms, 1990).

150

5.2.3.2 Moral Reasoning Outcomes

Five of the previous studies included a measure of moral reasoning, or a

measure of the cognitive distortions associated with immature moral reasoning,

providing mixed results. Whilst Jones (1991) reported small, statistically

insignificant improvements, another report found that both the intervention and

control groups improved (Gundersen & Svartdal, 2006). Three studies reported

no significant changes in moral reasoning. Again this is in stark contrast to the

current research which found that ART participants increased their moral

reasoning ability significantly, with a large effect size, whilst control participants

did not.

Several methodological differences between the studies may have contributed

to these contrasting findings. Three of the studies (Colman, Pfeiffer & Oakland,

1992; Currie et al 2009 and Glick & Goldstein, 1987) used samples which were

considerably older than those current research and as moral reasoning appears

to relate to age and ability (Gibbs, 2010: Guerra & Bradshaw, 2008), it could be

suggested that these samples experienced a ‘ceiling effect’ in relation to their

moral reasoning development. It could also be suggested that, as only one of

the studies included young people with differing levels of needs in the group

and this study found significant improvements in moral reasoning (Gundersen &

Svartdal, 2006) the number of ‘morally mature’ individuals required to support

and challenge the group with more principled reasoning was not sufficient,

leading to a lack of assimilation of more mature moral judgements (Dukerich et

al 1990; Gibbs, 2004).

The current findings are consistent with evaluations of similar discrete

interventions incorporating moral reasoning discussion groups (Arbuthnot &

Gordon, 1986; Dukerich et al. 1990; Gibbs et al 1984). Which together appear

to suggest that providing enriched opportunities of social-perspective taking

through moral discussion groups is effective in increasing young people’s moral

reasoning maturity.

In keeping with the theory proposed by Gibbs (Gibbs, Basinger & Fuller, 1992;

Glick & Gibbs, 2011) and research adopting samples of adolescent offenders

(Gregg, Gibbs & Basinger, 1994; Palmer & Hollin, 1998; Stams et al. 2006) the

151

sample of adolescents identified for intervention in the current study had

immature moral reasoning ability at pre-test, receiving on average Stage 2

ratings for their responses to moral situations.

However, in contrast to the model proposed by Gibbs (2004) and previous

research findings (Eisenberg et al 1991), there was no evidence of a

relationship between increased moral reasoning ability and greater pro-social

behaviour. Previous research suggests that translation of cognitive changes into

improvements in behaviour occurs over time, described as a ‘sleeper effect’

(Leeman, Gibbs & Fuller, 1993. p.290). Arbuthnot and Gordon (1986) found that

at 1 year follow up the improved sociomoral development of their intervention

participants had continued to increase and positive behaviour changes were

also persisting and diverging further from the control group. Similarly Leeman,

Gibbs and Fuller (1993) found no immediate effects in relation to gains in moral

reasoning ability from their intervention programme. However, after 1 year the

intervention group’s gains in moral judgement correlated negatively and

significantly with rates of recidivism, suggesting a long term effect. These

results highlight the importance of including follow up measures in evaluation

research into the ART programme and other interventions aiming to instigate

behaviour improvements through cognitive change.

5.3 Reflections on Supplementary Qualitative Findings

Previous research into the efficacy of ART has predominantly adopted an

experimental approach, focusing on gathering quantitative evidence of

outcomes. In order to gather some qualitative feedback, which would illuminate

the quantitative findings and provide the stakeholders, particularly the EPS, with

detail to inform future implementations of the intervention in school settings, the

researcher explored the participant’s and facilitator’s perceptions regarding the

ART sessions, led by the following question:

What are the views of those involved in the initial pilot of the ART

intervention sessions, in relation to programme implementation, contents

and effectiveness?

152

The following sections will now summarise the themes developed from the data

gathered and consider these themes in relation to previous research and theory

and the possible implications for future practice.

5.3.1 Summary of Themes Constructed

All respondents felt that the ART intervention had been effective in supporting

the group members. Positive outcomes for the participants, which were

attributed to participation in the programme included changes in social

behaviour which were not captured by the quantitative measures employed in

this study.

A range of positive outcomes were identified, from evidence of the explicit

taught skills being put into practice, changes in internal state and knowledge,

increased socially desirable observable behaviours and reports of positive

responses of others, as well as additional personal benefits for the individuals

such as feeling more mature and having a more positive view of their future.

With regard to the participant’s perceptions of the ART programme, several

salient themes and subthemes were evident in both the group member and

facilitator feedback. In regard to the content of the intervention, the

effectiveness of the games and role plays and the relevance and enjoyment of

the moral reasoning sessions were present in both types of data, suggesting

that the success of these components was noticed by both those conducting the

intervention and those participating. However, in relation to the implementation

of the programme, issues were highlighted at the organisational level including

difficulties with scheduling and several individuals indicated that careful

consideration of optimal group composition was of great importance. These two

areas will be discussed further in section 5.3.2.

Furthermore, in terms of the group member’s perceptions of the contents of the

sessions, themes arose which suggested that whilst the anger control sessions

were considered the least successful of the three intervention components, the

young people appreciated the strategies and teaching which were applicable to

the real world and half indicated that overall the sessions were enjoyable

experiences.

153

Other themes identified in the facilitator questionnaires suggested that whilst the

language used in the programme and their associated concepts were

sometimes too difficult for the participants to access, employing additional

strategies employed to tailor the sessions and motivate students, as well as

nurturing a positive environment, were all factors which were associated with

successful implementation of the programme and positive outcomes. A further

salient theme in the facilitator questionnaires encouraged the adoption of an

informed selection process, with specific characteristics highlighted as being

important for successful participation including motivation to change and their

difficulties being specifically related to aggression, as opposed to non-

compliance. The facilitators were also concerned that the participants should

have been provided the opportunity to volunteer for the intervention to increase

‘...buy in...’

5.3.2 Linking the Themes Constructed to Theory and Previous Research

Only two of the studies previously highlighted in Chapter 2 report qualitative

findings in relation to the impact of ART. Currie et al (2009) provided an

illustrative case study as part of their wider research into the implementation of

ART within Australian youth justice custodial settings. Facilitators were initially

concerned that Subject C was unable to follow the programme material.

However, later evidence suggested that he was able to apply the anger control

skills learnt during an inflammatory situation with a peer and staff at the centre

reported further observations of changes in his behaviour, including increased

avoidance of confrontations. These findings reflect the comments in the current

study which pertain to the subthemes ‘positive change noticed by others outside

of the sessions’, ‘lowered aggressive behaviour’ and ‘acquisition of explicitly

taught anger management techniques’. These similar findings might suggest

that these positive outcomes are common effects associated with the ART

programme, given the different settings and contrasting participants.

Interviews with six teachers, each trained as ART facilitators, also revealed

several interesting themes which closely align with those in the current research

(Sudbeck, 2010). Firstly all felt the intervention was useful in providing the

students with skills for everyday life, just as all of the facilitators felt that the

154

current implementation of ART had been effective for some of the individuals in

their groups. The majority of Sudbeck’s participants also identified the moral

reasoning component as being the most useful, in keeping with the views of

both the group members and facilitators in this study. Finally four out of the six

reported deviating from the programme to improve the applicability of the

contents to the student’s lives, just as the present facilitators described

differentiating the content to make it ‘relevant’ to the young people in their

group.

There was one theme which suggests that each group had differing

experiences of implementing the programme. Whilst all of Sudbeck’s facilitators

reported feeling supported by the staff and administration at their facility,

comments made by the EP facilitators suggested that organisational factors

such as a lack of recognition at a whole school level and external support were

a barrier to successful implementation and behaviour change. This highlights a

difficulty which is associated with external facilitators implementing programmes

in an unfamiliar setting. The implications of this will be discussed further below.

Several of the themes constructed in the current research align closely with the

organisational level factors identified as important in ensuring high quality

intervention implementation by researchers associated with the implementation

psychology movement (Denton, Vaughan & Fletcher, 2003; Domitrovich et al

2008; Fixsen et al 2005). Both the facilitators and participants also highlighted

elements of group composition which they felt contributed to the success of the

intervention. The issue of non voluntary participation was also frequently

referred to among the facilitator questionnaires. Subthemes from these three

more salient areas of data will now be explored in more detail.

5.3.2.1 Organisational Structures and Support

Need for recognition of the programme at a whole school level

One facilitator noted that a celebration should be included at the end of the

programme to recognise the pupil’s achievements, which should be given

importance at the whole school level. Another felt that future implementations

should ensure that senior management show a commitment to the project.

Gundersen, Finne and Olsen (2006) believe that administration have a key role

155

to play in the implementation of ART, acting as both a ‘driving force’ and

‘organiser’ of the programme and ART principles can ‘permeate the

organisation’s activities’ (p.56).

Administration communicate that interventions are a priority, making their

attendance during the planning and implementing phases highly important. This

may have alleviated the practical difficulties encountered (Domitrovich, Moore &

Greenberg, 2012), for example, regarding the use of pre-existing school merit

systems, which Facilitator B believed had impacted upon the student’s

motivation and interest in ART. Administration would also have control over the

scheduling of sessions, to ensure that competing activities would not influence

the participant’s motivation (Gundersen, Finne & Olsen, 2006) and finally the

room used for the intervention, which some students resented attending, due to

negative perceptions about the room’s purpose i.e. ‘I feel like a retard coming

here’.

Research has highlighted that administrative support within an organisation is a

significant factor in the successful implementation of new practices (Durlak &

Dupre, 2008; Kam, Greenberg & Walls, 2003; Klinger at al 2003; Tierney &

Dowd, 2000), including those who studied ART (Jones, 1991). Coleman,

Pfeiffer & Oakland (1992) suggested that total staff support is a prerequisite for

all social skills training, after their facilitators perceived the intervention as an

‘add on’ to the primary treatments offered, Whilst Jones (1991) noted that

participants had little encouragement to practise their skills as the teachers

were unaware of the programme.

Engagement of support outside of the sessions was required

This theme contained suggestions such as the incorporation of parent sessions

and the engagement of school staff to support the generalisation of skills

outside of the intervention context. Gundersen, Finne and Olsen (2006) state

that instructions about ART should be provided to the entire staff through

information days and parents should receive information regarding the current

skills being practised so that skills are reinforced outside of the sessions.

Scheduling issues meant that in the current implementation such

communication did not occur.

156

Researchers suggest that key stakeholders are involved early on, during the

planning stages, in order to increase acceptance, which is found to relate to

programme implementation (Domitrovich, Moore & Greenberg, 2012). Parent

engagement can help overcome barriers to successful intervention including

attendance at the sessions, changing their own behaviour and reinforcing their

child’s positive behaviours (Kazdin, 1987; Telford & Farrington, 1996). Studies

have also found that active parental involvement can increase the beneficial

effects of CBT interventions (Sofronoff, Attwood & Hinton, 2005).

5.3.2.2 Group Composition

Optimal Group Size

Group interventions with adolescents are often most effective when group sizes

remain small, with numbers lower than 8 children recommended (Coppock &

Dwivedi, 1993; Gresham & Elliott, 1993). In all of the previous research in ART

the group sizes adhered to the recommended guidelines of 4-8 individuals

(Gundersen et al. 2014). Group members from this study had contrasting views

about the size of the group. Participant F felt that bigger groups led to problem

behaviours and difficulty accessing the intervention, a common issue in large

groups due to the variety of personalities and disinhibition (Argyle, 1994),.

Another (Participant E) felt that non-attendance led to a small group, which

hindered the interactive activities.

In the current study School A chose to place 12 students in the group. The

facilitator questionnaire revealed that this group developed challenging

behaviours and experienced multiple exclusions, changing the group

composition week by week. Facilitator C was concerned that this meant the

remaining members ‘...possibly felt under threat of this for at least some of the

programme.’ Certainly this concern reflects theory surrounding group therapy.

From acceptance and belonging, group cohesion develops, where the group

represents a ‘...psychological whole.’ (Reid, 1987, p.176). Providing a safe

climate within the group is important. It provides relationship building

opportunities with peers and adults and consistency which may not be present

in their lives outside of the group (Malekoff, 2004). It is possible that threats to

157

this solidarity, safety and identity such as removing group members, could

impact upon this cohesion and in turn their effectiveness as a team.

Heterogeneity of need within the group is important

Facilitators suggested that there was a need for ‘...an equal balance of role

models with those who have immature anger control skills...’,(Facilitator D) ‘not

just the 8 ‘worst’ pupils in the school’ (Facilitator F). In keeping with a model

applied in previous research (Gundersen & Svardal, 2006/2010) and advocated

in the handbook (Gundersen, Finne & Olsen, 2006), both role models and target

pupils were identified during the selection process. To maintain ethical practice

the role models were also those with ‘low level’ needs who would benefit from

participation in the intervention.

However, due to unforeseen circumstances, an intended screening measure

was not employed and therefore professional judgements from staff at the

school were used for identification of these roles, with the support of the

researcher where possible. Descriptions provided by these staff members

(Please see Appendix XXIII) suggested that actually the target pupils and role

models selected had similar difficulties. For example, both contained individuals

described as ‘disruptive’ ‘rude’ and ‘verbally abusive’ as well as references to

physical fighting or aggression. They could therefore be considered a

homogenous group, such as those employed in Dishion’s research (Dishion &

Andrews, 1995; Dishion, McCord & Poulin, 1999). The iatrogenic effects noted

in these studies could therefore be the reason that the facilitators commented

that there was a need for ‘balance’ of difficulties and be behind the lack of

significant improvements in social behaviour for the intervention group. In future

the application of more stringent selection procedures would support the

identification of ‘suitable’ role models.

Role models need high status

In keeping with the subtheme above, the facilitators commented that choosing

older and higher status role models would be beneficial. Hierarchies within

groups are often influenced by the age and social status of the potential leaders

(Argyle, 1994). Social learning literature is clear that observational learning

158

depends upon interpersonal attraction and the existence of interesting qualities,

which enable some models to ‘...command greater attention than others’

(Bandura, 1977. p.6-7), which leads to increased emulation and imitation by

those present, setting the dominant trends in the social circle (Adler & Adler,

1998 cited in Lease, Musgrove and Axelrod, 2002). This power may stem from

an ability to control rewards and punishment, the individual’s perceived

competence or their likability among other things (Johnson & Johnson, 2009).

According to research age is also an important consideration, as older mentors

have been found to be successful in supporting younger children in improving

their interpersonal skills and avoiding anti-social behaviour (Dearden, 1998;

Gensemer, 2000; Sheehan et al. 1999). These observations, along with the

previous section, highlight the importance of careful selection of role models, as

those who are younger and less central in relation to their social centrality, may

not command the power required to influence the group positively. It is equally

important to considered composition at the ‘whole-group’ level. If the highest

status individuals are those who display challenging behaviour then it could be

suggested that the group may adopt these behaviours instead of those

displayed by the role models.

In future practice these ideas contribute significantly to the planning and

preparation stages of implementation. Not only in terms of a careful selection

process but Gensemer (2000), along with Facilitator F in the current study, also

highlighted the need to train the mentors in interpersonal skills and involve them

in preparation activities to ensure successful, informed support. Therefore in

future implementations it would be useful for those leading the programme to

meet with the role models prior to the sessions to enable them to understand

their role.

5.3.2.3 Non Voluntary Participation

The issue of ‘non-voluntary attendance’ appeared several times throughout the

questionnaire data. Facilitators felt that participants should ‘choose’ (Facilitator

F) or commented about ensuring ‘buy in’ in future whilst stating that they would

not force participation (Facilitator A). It could be suggested that some of the

group members, who were all selected onto the programme by a member of

159

school staff, did not feel they required support or were not ready to change the

behaviours focused upon in the intervention sessions. This would not only have

contributed to the high rates of attrition and non-significant findings but was also

hypothesised to be one of the causes of challenging behaviour faced by some

of the facilitators during the sessions. According to the stages of change model

devised by Prochaska, Norcross and DiClemente (1994) changes in behaviour

only start to occur when someone moves past precontemplation, in which they

deny a problem exists and resist change, and into contemplation and

preparation, whereby the problem is acknowledged and the individual begins to

make plans to take action towards solving it.

Whilst an entirely voluntary selection process may seem to contradict the

concerns of several other facilitators that participants must possess several

specific characteristics, future implementations may benefit from a introductory

meeting for the prospective group members, whereby the intervention is

advertised and attendance voluntary, followed by a more detailed selection

process based on ‘suitability’ for the programme. Alternatively some preparatory

work with ‘target individuals’ to help them to move from a state of

precontemplation to recognising the need to change their behaviours may also

improve their motivation to engage with the ART programme.

5.3.3 Summary

The quantitative results in the current study began to raise ethical concerns,

regarding the possibility that the programme had exaggerated some of the

intervention participant’s problem behaviours and social skills deficits in

accordance with parental report data. However, the comments received in the

qualitative data suggest otherwise, both facilitators and group members

consistently reported positive outcomes that they attributed to participation in

ART.

Themes were identified that could be easily incorporated into future practice in

relation to the programme’s contents and implementation, for example creating

a more detailed participant specification, adopting careful selection procedures

to ensure effective group composition, enlisting external support from parents or

160

teachers and preparing the implementation with those in administration to

ensure that organisational level barriers are dealt with in advance.

In light of the relationships noted between organisational support and

implementation success future investigations, possibly utilising an action

research approach, would be most beneficial to gain a greater insight into the

different elements of organisational support and the processes and structures

which help or hinder the administration and external personnel contributing to

the success of evidence-based programmes in school settings.

Whilst the results here suggest that small, heterogenous groups with high status

role models is the most effective composition for ART groups, further research

may wish to focus upon different characteristics of role models and the factors

mediating their influence over the ‘target’ individuals within ART, so that suitable

individuals are identified to support the programme.

5.4 Methodological Limitations

The differences noted between the outcomes reported in previous research and

the current findings may be due to procedural difficulties which will now be

discussed in detail. Conclusions drawn should be viewed tentatively in light of

these issues.

5.4.1 Issues of Internal Validity

Table 3.9 displayed several threats to internal validity. This section will focus

upon the threats which were considered to be particularly problematic in the

current research.

Firstly it would appear that there were differences between young people

chosen to be part of the groups in each setting. One noticeable issue is that the

pupils in the control group settings stemmed from a wider range of ages than

those in the intervention group. One school (School F) differed from the other 5

settings considerably, having lower levels of pupils with English as an Additional

Language and Free School meals as well as an older sample. Six scales on

the measures employed also revealed significant differences between the

control and intervention groups at pre-test, which suggests that initially the

control group had lower level needs than those in the intervention groups. This

161

is supported by the average standardised scores for each group which suggest

that the intervention group displayed behaviours which were outside of the

‘average’ range on five of the six standardised measures, whilst the control

group only received diminished social skills scores from self-report data and

above average problem behaviours when rated by teachers.

These factors may have contributed to the spontaneous improvements in

externalising and bullying behaviours noted between the pre and post-

measures in the control group. As previously suggested, being older and more

socially skilled may have meant that the pupils in the control group were able to

reflect upon their behaviour and make changes independently. The wait list

control group may also have felt the need to adopt compensatory strategies,

once alerted to their future involvement in the intervention and following their

participation in the pre-measures. Their heightened age and social competence

may have supported them in doing so.

Employing randomisation procedures and a matched pairs design, with

screening measures of the pupil’s social competence skills and behaviour

difficulties, would have eliminated these issues. However, both would require

the control and intervention conditions to be within the same setting, which

could lead to the treatment diffusion issues observed in previous research,

making it difficult to discriminate the impact of participation in the intervention. It

would also be difficult to match participants on their profiles of social skills

strengths and difficulties. Complexities such as these have lead researchers to

claim that RCTs are unfeasible within school contexts (Cohen, Manion &

Morrison, 2008; Greig, 2001)

‘History’ was a considerable threat in the current research, due to the use of six

separate settings, in which different members of staff rated the students and

different facilitators ran the groups. Whilst actions were taken to maintain

consistency across environments, such as standardised measures, instructions

and treatment fidelity checks, issues may still have remained which make it

difficult to attribute the outcomes measured solely to the effects of the

intervention.

162

Differences were noted within each individual environment, such as the level of

support provided for the intervention in the wider school context and levels of

support for the facilitators within the group sessions. The teacher-raters may

have differed in their mood, the amount and type of contact with the pupil and

previous relationship with the pupil they were rating. Finally, the facilitators

employed different forms of reinforcement and modes of delivery, for example

School B included additional role models 3 weeks into the programme and

School C used a gift card reward system. They may also have had different

presenting styles and opportunities to build relationships with the young people.

All of these factors could have influenced the results gathered. Again utilising

the same settings for the control and intervention groups, as well as the same

raters and facilitators, would have provided additional levels of control but this

was not feasible in the current research.

It may be that the anger management and social competence strategies taught

as part of ART required more time to produce positive outcomes. The previous

studies that have included follow up measures found that the positive outcomes

were long lasting (Glick & Goldstein, 1987) and the improvements in problem

behaviours displayed were even more apparent during the post-test period then

they had been over the course of the intervention (Langeveld, Gundersen &

Svartdal, 2012). Doctoral timescales meant that the intended Time 3 measure

was not possible. However, the lack of research into the long term effects of

ART suggest that longitudinal research is needed.

There were large rates of attrition in the research, despite collaboration with

schools prior to the intervention to prevent such issues. At post-test 56% of the

original sample remained, 83% of the original control group and only 35% of the

intervention group. School B withdrew from the project entirely after 4 weeks as

they felt that the students were not engaging with the programme. Analyses

suggested that there were no differences noted on the measures between those

who left and those who remained (described in section 4.1.2). However, the

attrition may have lead to a biased sample, for example the intervention

participants who remained may have been coerced by the school to stay, which

could have had an effect on their motivation to engage with the programme.

The attrition also meant that the sample that remained was much smaller than

163

at pre-test, which lowered the statistical power, increasing the chance of a Type

II error and limiting conclusions.

5.4.2 Issues of External Validity

The use of a small, homogenous sample, which did not use screening

measures for identification nor randomisation to conditions, suggests that

attempts to apply the conclusions drawn to the wider population would be

invalid. Instead the intention was to be ‘...interested in a specific finding in its

own right...’ (Robson, 2011, p.91) and provide an indication of the impact of the

initial pilot of the intervention within one local authority setting, alongside

additional qualitative data to guide future practice. It is also hoped that the

findings will also contribute to the existing evidence base and stimulate further

research, as this is the first investigation into the effects of ART with a school

based population within the United Kingdom.

5.4.3 Issues of Reliability

Some features of the measures adopted in the current study may have led to

issues concerning the reliability of the findings gathered. Firstly the SSIS-RS

was standardised using a sample of American children. This suggests that the

standardised scores referred to in this research should be viewed with caution,

as the current sample were from the UK.

Secondly, there were some difficulties experienced during the study regarding

the accessibility of the measures. Whilst parents were offered support from the

settings to complete the measures together, some did not attend, leading to

attrition in the parental data. Several pupils also found the SRM-SF difficult to

access, despite one to one support, which may have led to the high number of

‘unscorable’ responses, and in turn, the small sample of data regarding pupil’s

moral reasoning maturity.

The SRM-SF also attained poor inter-rater reliability scores in the current study,

in relation to agreement between raters of the exact global stage given to each

questionnaire, which may suggest that the significant improvements noted, are

due to inconsistent scoring. However, it is important to note that the second

rater did not undertake the prescribed self-training that the SRM-SF handbook

164

provides. This may have influenced their ability to rate the questionnaires

accurately.

A number of these issues could have been overcome by employing simpler,

shorter measures of behaviour and a recognition measure of moral reasoning

as opposed to a production measure. However, the current measures were

chosen because of their use in previous research into ART, making the results

comparable, and the increased detail that they gathered about the concepts that

they measured.

Finally self-report measures are often criticised for being vulnerable to demand

characteristics. In the current research it could be suggested that the group

members responded in a socially desirable manner, not wanting to be honest

about the extent of their behaviour difficulties or they may not have the ‘...self-

reflective thought...’ needed to answer accurately (Feindler & Baker, 2004,

p.36). This would explain why the positive improvements recorded for the

control group were only present in their self-report data. Actions were taken to

compensate, for example parents and teachers also completed comparable

measures, avoiding reliance on one data source for a more complete picture

(Renk & Phares, 2004) and providing information about generalisation of

behaviours noted across several contexts. Future research may consider

triangulation with measures such as direct observation in naturalistic contexts,

which are considered to be more sensitive to short term changes. However,

they too are not without criticism as they lack theoretical grounding (Gresham et

al. 2004).

5.4.4 Reflections on the Challenges Encountered in Real World Research

It is widely acknowledged that conducting research in real world settings,

particularly the field of education, is complex. Researchers have identified many

challenges which can hinder the research process and in turn the validity of the

conclusions drawn (Mertens, 2010; Robson, 2011) such as differing agendas,

sampling bias and poor communication. The current research experienced

several instances of such difficulties. For example, teaching staff were often

reluctant for the students to miss lessons to attend the ART sessions and pre-

existing school systems sometimes made communication with parents difficult.

165

Whilst the researcher maintains that the quasi-experimental group design

employed currently was the most appropriate given the purpose of the research

and the questions posed (Cohen, Manion & Morrison, 2008; Slavin, 2002), the

challenges encountered and their effects on the ‘robustness’ of the conclusions

drawn highlight the value of more descriptive designs, such as single-case

experiments (Mertens, 2010) and those which incorporate an element of

collaboration between the researcher and professional practitioners, such as

action research (Gray, 2014).

5.5 Implications of the Findings

The methodology used in this study has provided an exploration into the

implementation and impact of ART within real world settings. Although it is not

possible to generalise to other settings due to the small sample and

methodological limitations outlined previously, it is possible to suggest areas

which would benefit from further investigation and factors identified in the

current study which may be applicable to successful implementation in other

contexts.

5.5.1 Implications for Future Research

Further replications of the current study are required in order to broaden the

evidence base regarding the efficacy of ART with adolescent samples within UK

contexts. It would be beneficial to gain an insight into factors which mediate the

effects of the intervention in UK school environments, including the gender, age

and cognitive ability of the participants; the possible differential effects of

utilising facilitators internal and external to the environment and the

characteristics of the role models which influenced their power over the group.

Such research would provide an opportunity to explore previously mentioned

hypotheses regarding the cultural connectedness of the intervention and in turn,

modifications could be made to make ART more culturally relevant.

The trends noted in the parent report data warrant further investigation, given

the paucity of ART research that has previously included parental measures, in

order to see whether ART interventions can influence parental perceptions of

their child’s problem behaviours in the home environment. It may be that such

166

studies focus upon the mechanisms that support the generalisation of skills

outside of the school context.

In regard to methodology, there were several issues within the current design

which could inform future research. Firstly longitudinal measures are highly

important, particularly in light of the hypothesis that changes in moral reasoning

maturity precede associated behaviour modifications. Employing a comparison

group would improve internal validity as, in the current study, effects noted

could be due to the raised attention and other factors associated with taking

part in intervention sessions which the control group did not have.

Larger samples, from the same settings, identified using valid screening

measures and randomised and matched into the different conditions would also

allow the researcher to draw more robust conclusions and generalise to wider

populations. However, it would be advisable to ensure measures are taken to

avoid secondary diffusion, or such effects could complicate the researcher’s

ability to partial out the effects of the intervention.

Action research is flexible in nature and focuses upon ‘addressing real world

problems’ (Gray, 2014. Pp.328) and improving conditions (Robson, 2011),

making it well suited to research conducted within complex, dynamic

educational environments. This design would be highly appropriate as it would

enable the researcher to capture detail at the planning and preparation stage as

well as richer information regarding factors thought to impact upon effective

implementation such as collaboration with those within the organisational

settings and enlisting external sources of assistance such as parents, teachers

and peers. This is due to action research providing opportunities to capture

different forms of data, as they evolve at different points during the

implementation process, whilst maintaining an evaluative element from which

changes could be fedback to improve the implementation. The collaborative

nature of action research may also be beneficial, as attempting to impose

certain conditions as an external researcher can be challenging, for example in

the current project it became difficult to ensure similarity in the rewards,

involvement of staff in the sessions and participant’s choice to attend between

the different settings. The vast discrepancy between what was considered

167

appropriate by the staff at the different schools illustrated the importance

tailoring the intervention and implementation to the environment, the pre-

existing systems and the individuals within it. Collaborating with the settings

about these decisions may have been a better method of ensuring some

consistency and may ultimately have impacted upon the outcomes observed.

Further qualitative research would also be beneficial, as this is the first study to

consider group member’s perceptions of the programme. Feedback can support

the modification of ART to improve the implementation and outcomes. Possible

further sources of information include parents and teachers, who can contribute

to discussions surrounding generalisation of skills.

5.5.2 Implications for Practice

This research has highlighted some factors which the EPS and other ART

practitioners may wish to consider when implementing ART within school

settings.

Firstly the qualitative data allowed for many suggestions of improvements,

which the current participants attributed to the success of the intervention, that

could easily be incorporated into future implementations of ART including

participant characteristics which could be adopted as selection criteria,

particularly when selecting role models who require status within the group in

order to assert the positive influence desired, adaption of the contents to make

it more accessible and relevant and the importance of nurturing a positive

environment.

The qualitative data also highlighted the significance of the preparation stage in

ensuring successful implementation. Several different elements were identified

including communication and collaboration with different levels within the

organisation to ensure that practical needs and support systems were in place

and recruiting sources of external support such as teachers, parents and peers.

EPs are also well placed to contribute to the selection process, conducting valid

and reliable measures that will support the identification of group members with

‘suitable’ characteristics and ensuring optimal group composition (Rait, Monsen

& Squires, 2010).

168

Several practical implications arise from the data regarding moral reasoning

development. Firstly, whilst many programmes aimed at promoting social skills

focus on changing observable behaviours, these results highlight the

importance of including elements within the intervention package which nurture

cognitive skills associated with positive behaviour change such as moral

reasoning. As these have been found to not only be responsive to intervention,

but have previously been associated with greater prosocial behaviour

(Eisenberg et al. 1991) and may in fact be a prerequisite for more permanent

behaviour change in the case of the ‘sleeper effect’ hypothesis (Leeman, Gibbs

& Fuller, 1993). This also highlights the importance of monitoring the impact of

interventions conducted in schools over prolonged periods of time, as more

noticeable changes may occur weeks or months after the programme has

concluded. This study also utilised measures of both cognitive concepts and

observable behaviours, something which could be incorporated into educational

psychology practice and would provide richer evidence in relation to programme

effectiveness.

In relation to evidence-based practice the non-significant results observed in the

‘problem behaviour’ and ‘social skills’ variables highlight that potential practical

and ethical issues can arise when implementing an intervention package,

despite having high treatment integrity. Careful consideration is required in

relation to the mechanisms through which changes are expected to occur, the

suitability of the context in which it is to be implemented and the measures

utilised or type of evidence which will be gathered. It could be suggested that

the positive responses identified during the qualitative data analysis highlight

the importance of different forms of data collection in applied contexts, which

could be applied to both evidence based practice endeavours and the wider

evaluation of Educational Psychologist’s case work.

In keeping with this theme the current study adopted a multi-informant process,

utilising several dependent variables, to ensure a more comprehensive

evaluation. Alongside this pupil’s and facilitator’s views were gathered to

improve and refine future experiences of the programme. These principles

could also be employed in wider educational psychology practice, when

attempting to evidence the impact of many forms of EP involvement.

169

Finally, the deterioration of reported outcomes in the parental data suggests

that support for the families of young people with social and behavioural

difficulties is vitally important. Given the research evidence regarding the

increased positive outcomes when parents are involved in the programme

(Sofronoff, Attwood & Hinton, 2005), this is an area in which could have a

considerable impact in relation to intervention success.

5.6 Conclusions

5.6.1 Unique Contribution of the Current Research

The primary unique contribution of the current research to the existing evidence

base was investigating the impact of ART sessions when conducted within

mainstream secondary school settings within the UK. The research design

employed provided opportunities to consider quantifiable changes in behaviour

and cognition, whilst also gaining an understanding of the perspectives of those

involved in this initial pilot, in relation to factors influencing the success of the

programme. These findings can potentially be used in future applications of

ART and guide further research.

The literature reviewed in section 2.6 also indicated that much of the existing

research with adolescent samples had been conducted outside of school

settings, using measures closely aligned with the programme contents, with

only one or two sources of reporting and facilitators who were internal to the

setting.

This study explored the success of ART when implemented in a more

‘preventative’ manner. The sample of young people perceived as displaying

social skills and behaviour difficulties attended mainstream settings. Newly

trained professionals from the EPS, external to the school setting, facilitated the

sessions. According to Squires (2001) EPs have a vital role to play in working

preventatively, supporting the planning and implementation of social inclusion

interventions in UK schools (Denham et al. 2006). In order to gain further

information regarding the generalisability of the skills gained, the measures

employed were chosen because they represented the wider concepts which

ART aims to instil and three sources of evidence (parents, teachers and self-

170

report) supported the investigation of the application of these new skills across

environments.

5.6.2 Summary of Findings

This research presents the first quasi-experimental investigation into the impact

of ART, conducted within UK secondary school settings by recently trained

facilitators from the local EPS.

The lack of quantitative evidence of improvements in social skills and problem

behaviours for the intervention group has led to the conclusion that, on this

occasion, ART was unsuccessful in supporting the development of social skills

and alleviation of problem behaviours in the short term. The researcher

provides several possible explanations for this lack of positive outcomes,

including the programme contents being culturally bound, the measures

representing more generalised, wider concepts than those typically used in ART

research and a lack of generalisation of skills to the wider environment. The

influence of methodological limitations including threats to internal validity and

reliability and problems that arose during the implementation of the programme

such as poor attendance and high attrition were also considered.

However, the intervention group did ascertain a large positive change in their

moral reasoning maturity from pre to post-measure, whilst the control group,

who attended normal lessons, did not change significantly. This change was

explained in terms of a possible ‘sleeper effect’ (Leeman, Gibbs & Fuller, 1993,

p.290) whereby changes in cognition precede associated positive changes in

behaviour. In order to confirm this hypothesis, longitudinal research is required,

to investigate whether changes in observable behaviour do indeed develop at a

later stage.

The qualitative data offered a contrasting perspective, whereby all of those

involved in this initial pilot of ART felt that the intervention had been effective,

attributing a range of positive outcomes such as group member’s knowledge of

anger control strategies and social skills to their involvement in the programme.

The qualitative data also gave rise to several suggestions regarding the

contents and implementation of the sessions considered to contribute to the

171

success of ART. Some of these suggestions had considerable implications for

the planning stage of intervention implementation, such as collaboration with

the organisation to enlist the support of administration and external sources of

reinforcement including teachers and parents.

172

6. References

Arbuthnot, J., & Gordon, D. A. (1986). Behavioral and Cognitive Effects of a Moral Reasoning Development Intervention for High-Risk Behavior-Disordered Adolescents. Journal of Consulting and Clinical Psychology, 54(2), 208–16.

Arcia, E. (2006). Achievement and Enrolment Status of Suspended Students: Outcomes in a Large, Multicultural School District. Education and Urban Society, 38(3), 359–369.

Argyle, M. (1994). The Psychology of Interpersonal Behaviour. London: Penguin.

Argyle, M., & Kendon, A. (1967). The Experimental Analysis of Social Performance. In Berkowitz, L. (Ed.), Advances in Experimental Social Psychology. New York: Academic Press.

ATL. (2012). Association of Teachers and Lecturers Press Release. Retrieved from http://www.atl.org.uk/Images/27 for 30 March 2012 - pupil behaviour survey FINAL.pdf

Attwood, G., Croll, P., & Hamilton, J. (2003). Re-engaging with education. Research Papers in Education, 18(1), 75–95.

Babbie, E. (2010). The Practice of Social Research. California: Wadsworth.

Bandura, A. (1969). Principles of Behaviour Modification. London: Holt, Rinehart and Winston.

Bandura, A. (1973). Aggression- A Social Learning Analysis. New Jersey: Prentice-Hall.

Bandura, A. (1977). Social Learning Theory. New Jersey: Prentice-Hall.

Bandura, A., Ross, D., & Ross, S. A. (1963). Imitation of Film-Mediated Aggressive Models. The Journal of Abnormal and Social Psychology, 66(1), 3–11.

Barnoski, R., & Aos, S. (2004). Outcome Evaluation of Washington State’s Research-Based Programs for Juvenile Offenders. Washington: Washington State Institute for Public Policy.

Barriga, A. Q., Landau, J. R., Stinson, B. L., Liau, a. K., & Gibbs, J. C. (2000). Cognitive Distortion and Problem Behaviors in Adolescents. Criminal Justice and Behavior, 27(1), 36–56.

173

Barriga, A. Q., Morrison, E. M., Liau, A. K., & Gibbs, J. C. (2001). Moral Cognition: Explaining the Gender Difference in Antisocial Behavior. Merrill-Palmer Quarterly, 47(4), 532–562.

Basinger, K., Gibbs, J., & Fuller, D. (1995). Context and the Measurement of Moral Judgement. International Journal of Behavioral Development, 18(3), 537–556.

Beck, A. T. (1991). Cognitive Therapy and the Emotional Disorders. London: Penguin.

Beck, A. T. (1999). Prisoners of Hate. New York: Harper Collins.

Beck, R., & Fernandez, E. (1998). Cognitive-Behavioral Therapy in the Treatment of Anger : A Meta-Analysis. Cognitive Therapy and Research, 22(1), 63–74.

Beelman, A., Pfingsten, U., & Losel, F. (1994). Effects of Training Social Competence in Children: A Meta-Analysis of Recent Evaluation Studies. Journal of Clinical Child Psychology, 23(3), 260–271.

Biesta, G. (2007). Why “What Works” Wont Work : Evidence-Based Practice and the Democratic Deficit in Educational Research. Educational Theory, 57(1), 1–22.

Bjorkqvist, K., Lagerspetz, K. M. J., & Kaukiainen, A. (1992). Do Girls Manipulate and Boys Fight ? Developmental Trends in Regard to Direct and Indirect Aggression. Aggressive, 18, 117–127.

Blase, K. A., Van Dyke, M., Fixsen, D., & Wallace Bailey, F. (2012). Implementation Science. In Kelly, B. & Perkins, D (Eds.), Implementation Science. New York: Cambridge University Press.

Bowlby, J. (1969). Attachment and Loss Volume 1. London: Hogarth Press.

Boyatzis, R. E. (1998). Transforming Qualitative Information: Thematic Analysis and Code Development. California: Sage.

BPS. (2005). Inclusive Education Position Paper. Leicester: BPS.

BPS. (2010). Code of Human Research Ethics. Leicester: British Psychological Society.

Braun, V., & Clarke, V. (2006). Using Thematic Analysis in Psychology. Qualitative Research in Psychology, 3(2), 77–101.

Braun, V., & Clarke, V. (2014). Successful Qualitative Research. London: Sage.

Breakwell, G. M. (1999). Coping With Aggressive Behaviour. Leicester: British Psychological Society.

174

Brennan, P., Mednick, S., & Kandel, E. (1991). Congenital Determinants of Violent and Property Offending. In Pepler, D. & Rubin, K. (Eds.), Development and Treatment of Childhood Aggression. New Jersey: Lawrence Erlbaum.

Bronfenbrenner, U. (1996) The Ecology of Human Development. London: Open University Press.

Brown, J., & Winterton, M. (2010). Violence in UK Schools : What is Really Happening. Cheshire: British Educational Research Association.

Brunton, G., Stansfield, C., & Thomas, J. (2012). Finding Relevant Studies. In Gough, A., Oliver, S., & Thomas, J. (Eds.), An Introduction to Systematic Reviews. London: Sage.

Buchner, A., Erdfelder, E., Faul, F., & Lang, A.-G. (2014). G*Power version 3.1.9.2. Retrieved February 06, 2015, from http://www.gpower.hhu.de/

Cairns, R. B., Cairns, B. D., Neckerman, H. J., Gest, S. D., & Gariepy, J.-L. (1988). Social networks and aggressive behavior: Peer support or peer rejection? Developmental Psychology, 24(6), 815–823.

Caldwell, M. F., & Van Rybroek, G. (2013). Effective Treatment Programs for Violent Adolescents: Programmatic Challenges and Promising Features. Aggression and Violent Behavior, 18(5), 571–578.

Cartledge, G., & Milburn, J. (1995). Teaching Social Skills to Children and Youth. Massachusetts: Allyn and Bacon.

Caspi, A., Henry, B., McGee, R., T, M., & Silva, P. (1995). Temperamental Origins of child and adolescent behaviour problems. Child Development, 66(1), 55–68.

Chalmers, I. (2003). Trying to do More Good than Harm in Policy and Practice: The Role of Rigorous, Transparent, Up-to-Date Evaluations. The Annals of the American Academy of Political and Social Science, 589(1), 22–40.

Clarke, D. (2003). Pro-Social and Anti-Social Behaviour. East Sussex: Routledge.

Cline, T. (1992) The Assessment of Special Educational Needs. An

International Perspective. London: Routledge.

Cohen, J. (1992). A Power Primer. Psychological Bulletin, 112(1), 155–159.

Cohen, L., Manion, L., & Morrison, K. (2008). Research Methods in Education. Oxon: Routledge.

175

Colby, A., Kohlberg, L., Gibbs, J., & Lieberman, M. (1983). A Longitudinal Study of Moral Judgment. Monographs of the Society for Research in Child Development, 48(1-2), 1–124.

Coleman, M., Pfeiffer, S., & Oakland, T. (1992). Aggression Replacement Training with Behaviorally Disordered Adolescents. Behavioral Disorders, 18(1), 54–66.

Cook, T. D. (2002). Randomized Experiments in Educational Policy Research: A Critical Examination of the Reasons the Educational Evaluation Community has Offered for not Doing Them. Educational Evaluation and Policy Analysis, 24(3), 175–199.

Cook, T. D., Shadish, W. R., & Wong, V. C. (2008). Three Conditions under Which Experiments and Observational Studies Produce Comparable Causal Estimates : New Findings from Within-Study Comparisons Abstract. Journal of Policy Analysis and Management, 27(4), 724–750.

Cooper, P., Smith, C. J., & Upton, G. (1995). Emotional and Behavioural Difficulties. London: Routledge.

Coppock, C., & Dwivedi, K. (1993). Group Work in Schools. In Dwivedi, K. (Ed) Group Work with Children and Adolescents. London: Jessica Kingsley.

Cornell, D., Peterson, C., & Richards, H. (1999). Anger as a Predictor of Aggression Among Incarcerated Adolescents. Journal of Consulting and Clinical Psychology, 67(1), 108–115.

Costenbader, V., & Markson, S. (1998). School Suspension : A Study with Secondary School Students. Journal of School Psychology, 36(1), 59–82.

Crick, N. R. (1996). The Role of Over Aggression, Relational Aggression, and Prosocial Behavior in the Prediction of Children’s Future Social Adjustment. Child Development, 67, 2317–2327.

Crick, N. R., & Dodge, K. A. (1994). A Review and Reformulation of Social Information-Processing Mechanisms in Children’s Social Adjustment. Psychological Bulletin, 115(1), 74–101.

Crick, N. R., & Dodge, K. A. (1996). Social Information-Processing Mechanisms in Reactive and Proactive Aggression. Child Development, 67, 993–1002.

Crosby, J. W. (2011). Test Review: F. M. Gresham & S. N. Elliott Social Skills Improvement System Rating Scales. Minneapolis, MN: NCS Pearson, 2008. Journal of Psychoeducational Assessment, 29(3), 292–296.

Csikszentmihalyi, M., & Larson, R. (1984). Being Adolescent. New York: Basic Books.

176

Currie, M. R, Wood, C. E., Williams, B., & Bates, G. W. (2009). Aggression Replacement Training in Australia: Youth Justice Pilot Study. Psychiatry, Psychology and Law, 16(3), 413–426.

Currie, M. R., Wood, C. E., Williams, B., & Bates, G. W. (2012). Aggression Replacement Training (ART) in Australia: A Longitudinal Youth Justice Evaluation. Psychiatry, Psychology and the Law, 19(4), 577–604.

Curulla, V. L. (1991). Aggression Replacement Training in the Community for Adult Learning Disabled Offenders. Unpublished doctoral dissertation. University of Washington, Washington. United States of America

Damon, W. (1999). The Moral Development of Children. Scientific American, 72–78.

Dancey, C. P., & Reidy, J. (2011). Statistics Without Maths for Psychology. Essex: Pearson Education.

Daniels, H., Cole, T., Sellman, E., Sutton, J., Visser, J., & Bedward, J. (2003). Study of Young People Permanently Excluded From School. Nottingham: HMSO

DCSF. (2007). Aiming high for young people: a ten year strategy for positive activities. Norwich: DCSF

DCSF. (2008). Improving behaviour and attendance: guidance on exclusion from schools and Pupil Referral Units. Nottingham: DCSF.

Dearden, J. (1998). Cross-age Peer Mentoring in Action. Educational Psychology in Practice, 13(4), 250–257.

Deffenbacher, J., Thwaites, G., Wallace, T., & Oetting, E. (1994). Social Skills and Cognitive-Relaxation Approaches to General Anger Reduction. Journal of Couselling Psychology, 41(3), 386–396.

Denham, A., Hatfield, S., Smethurst, N., Tan, E., & Tribe, C. (2006). The Effect of Social Skills Interventions in the Primary School. Educational Psychology in Practice, 22(1), 33–51.

Denton, C. A., Vaughn, S., & Fletcher, J. M. (2003). Bringing Research-Based Practice in Reading Intervention to Scale. Learning Disabilities Research and Practice, 18(3), 201–211.

DES. (1978). Special Educational Needs Report of the Committee of Enquiry into the Education of Handicapped Children and Young People (Report of the Warnock Committee). London: HMSO.

DfE. (1994). Pupils With Problems. London: DfE.

177

DfE. (2011a). Positive for Youth: Adolescent Brain Development. Nottingham: DfE.

DfE. (2011b). Positive for Youth. Nottingham: DfE.

DfE. (2012). Pupil Behaviour in Schools in England. Nottingham: DfE.

DfE. (2013a). Children with Special Educational Needs 2013: An Analysis. London: DfE.

DfE. (2013b). Statistical First Release. Permanent and Fixed Period Exclusions from Schools and Exclusion Appeals in England, 2011/2012. London: DfE.

DfE (2015) Special Educational Needs and Disability Code of Practice: 0-25 Years. London: DfE.

DfEE. (1997). Excellence for All Children. Suffolk: DfEE.

DfES. (2003). Every Child Matters. Nottingham: DfES.

DfES. (2005). Learning Behaviour. The Report of the Practitioners’ Group on School Behaviour and Discipline. Nottingham: DfES.

Dishion, T. J., & Andrews, D. W. (1995). Preventing Escalation in Problem Behaviors With High-Risk Young Adolescents : Immediate and 1-Year Outcomes. Journal of Consulting and Clinical Psychology, 63(4), 538–548.

Dishion, T., McCord, J., & Poulin, F. (1999). When Interventions Harm. American Psychologist1, 54(9), 755–764.

Domitrovich, C., Bradshaw, C., Poduska, J. M., Hoagwood, K., Buckley, J. A., Olin, S., Ialongo, N. (2008). Maximising the Implementation Quality of Evidence-Based Preventive Interventions in Schools: A Conceptual Framework. Advances in School Mental Health Promotion, 1(3), 6–28.

Domitrovich, C., Moore, J., & Greenberg, M. (2012). Maximising Social-Emotional Interventions. In Kelly, B. & Perkins, D. (Eds.), Handbook of Implementation Science for Psychology in Education. New York: Cambridge University Press.

Dukerich, J., Nichols, M., Elm, D., & Vollrath, D. (1990). Moral Reasoning in Groups: Leaders Make a Difference. Human Relations, 43(5), 473–493.

Dunsmuir, S., Brown, E., Iyadurai, S., & Monsen, J. (2009). Evidence Based Practice and Evaluation: From Insight to Impact. Educational Psychology in Practice, 25(1), 53–70.

Dunsmuir, S., & Iyaduria, S. (2007). Cognitive Behavioural Therapy: Effectiveness, Expertise and Ethics. DECP Debate, (122), 15–20.

178

Durlak, J. A, & DuPre, E. P. (2008). Implementation Matters: A Review of Research on the Influence of Implementation on Program Outcomes and the Factors Affecting Implementation. American Journal of Community Psychology, 41(3-4), 327–50.

Durlak, J. A, Fuhrman, T., & Lampman, C. (1991). Effectiveness of Cognitive-Behavior Therapy for Maladapting Children: a Meta-analysis. Psychological Bulletin, 110(2), 204–14.

Dwivedi, K., & Gupta, A. (2000). “Keeping Cool”: Anger Management Through Group Work. Support for Learning, 15(2), 76–81.

Edmondson, C. & Conger, J. (1996) A Review of Treatment Efficacy for Individuals With Anger Problems. Clinical Psychology Review. 16(3) 251-275

Eisenberg, N., Miller, P., Shell, R., McNally, S., & Shea, C. (1991). Prosocial Development in Adolescence: a Longitudinal Study. Developmental Psychology, 27(5), 849–857.

Ellis, A. (1975). Reason and Emotion in Psychotherapy. New Jersey: Lyle Stuart.

Engels, R., Finkenauer, C., Meeus, W., & Dekovic, M. (2001). Parental Attachment and Adolescents’ Adjustment: The Associations With Social Skills and Relational Competence. Journal of Counseling Psychology, 48(4), 428–439.

EPPI. (2007). Methods for Conducting Systematic Reviews. London: Institute of Education.

Erickson, J. A. (2013). The Efficacy of Aggression Replacement Training with Female Juvenile Offenders in a Residential Commitment Program. Unpublished Doctoral Dissertation, University of South Florida, Florida. United States of America.

Farrington, D. P. (1995). The Development of Offending and Antisocial Behaviour from Childhood: Key Findings from the Cambridge Study in Delinquent Development. Journal of Child Psychology and Psychiatry, 360(6), 929–964.

Faupel, A. (2002). Dealing With Anger and Aggression. In P. Gray (Ed.), Working With Emotions. London: Routledge Falmer.

Faupel, A., Herrick, E., & Sharp, P. (2011). Anger Management A Practical Guide. London: Routledge.

Feilzer, Y. M. (2010). Doing Mixed Methods Research Pragmatically: Implications for the Rediscovery of Pragmatism as a Research Paradigm. Journal of Mixed Methods Research, 4(1), 6–16.

179

Feindler, E. (1995). Ideal Treatment Package for Children and Adolescents with Anger Disorders. Issues in Comprehensive Paediatric Nursing, 18(3), 233–260.

Feindler, E., & Baker, K. (2004). Current Issues in Anger Management Interventions with Youth. In Goldstein, A. Nensen, R. Daleflod, B. & Kalt. M. (Eds.), New Perspectives on Aggression Replacement Training. West Sussex: John Wiley and Sons.

Feindler, E., & Ecton, R. (1986). Adolescent Anger Control. Oxford: Pergamon Press.

Feindler, E., Marriott, S., & Iwata, M. (1984). Group Anger Control Training for Junior High School Delinquents. Cognitive Therapy and Research1, 8(3), 299–311.

Felner, R. D., Favazza, A., Shim, M., Brand, S., Gu, K., & Noonan, N. (2001). Whole School Improvement and Restructuring as Prevention and Promotion. Lessons from STEP and the Project on High Performance Learning Communities. Journal of School Psychology, 39(2), 177–202.

Felson, R. B. & Tedeschi, J. T. (1995) Aggression and Violence: Social Interactionist Perspectives. Washington: American Psychological Association.

Ferguson, N., McLernon, F., & Cairns, E. (1994). The Sociomoral Reflection Measure-Short Form: An Examination of its Reliability and Validity in a Northern Irish Setting. British Journal of Educational Psychology, 64, 483–489.

Field, A. (2009). Discovering Statistics Using SPSS. London: Sage.

Field, A., & Hole, G. (2006). How to Design and Report Experiments. London: Sage.

Fien, J. (2002). Advancing Sustainability in Higher Education: Issues and Opportunities for Research. Higher Education Policy, 15, 143–152.

Fixsen, D. L., Blase, K. a., Naoom, S. F., & Wallace, F. (2009). Core Implementation Components. Research on Social Work Practice, 19(5), 531–540.

Fixsen, D., Naoom, S. F., Blase, K. A., Friedman, R. M., & Wallace, F. (2005). Implementation Psychology: A Synthesis of the Research : Florida: University of South Florida.

Fox, M. (2003). Opening Pandora’s Box: Evidence-based Practice for Educational Psychologists. Educational Psychology in Practice, 19(2), 91–102.

180

Fraser, M. W., Galinsky, M. J., Smokowski, P. R., Day, S. H., Terzian, M. a, Rose, R. a, & Guo, S. (2005). Social Information-Processing Skills Training to Promote Social Competence and Prevent Aggressive Behavior in the Third Grade. Journal of Consulting and Clinical Psychology, 73(6), 1045–1055.

Frederickson, N. (2002). Evidence-Based Practice and Educational Psychology. Educational and Child Psychology, 19(3), 96–111.

Frederickson, N., & Cline, T. (2009). Special Educational Needs, Inclusion and Diversity. Berkshire: Open University Press.

Frederickson, N., & Simms, J. (1990). Teaching Social Skills to Children: Towards and Integrated Approach. Educational and Child Psychology, 7(1), 5–17.

Fritz, C. O., Morris, P. E., & Richler, J. J. (2012). Effect Size Estimates: Current Use, Calculations, and Interpretation. Journal of Experimental Psychology: General, 141(1), 2–18.

Fuggle, P., Dunsmuir, S., & Curry, V. (2013). CBT with Children, Young People and Families. London: Sage.

Gable, R. A., Bullock, L. M., & Harader, D. L. (1995). Schools in Transition. Preventing School Failure, 39(3), 29–34.

Gansle, K. A. (2005). The effectiveness of school-based anger interventions and programs : A meta-analysis. Journal of School Psychology, 43, 321–341.

Gensemer, P. (2000). Effectiveness of Cross-Age and Peer Mentoring Programs. Information Analyses, 30, 1–15.

Gersch, I. S., & Nolan, A. (1994). Exclusions: What the Children Think. Educational Psychology in Practice, 10(1), 35–45.

Gibbs, J. C. (2004). Moral Reasoning Training. In Goldstein, A. Nensen, R. Daleflod, & B. Kalt, M. (Eds.), New Perspectives on Aggression Replacement Training. West Sussex: John Wiley and Sons.

Gibbs, J. C. (2010). Moral Development and Reality. Boston: Pearson Education.

Gibbs, J. C., Arnold, K. D., Ahlborn, H. H., & Cheesman, F. L. (1984). Facilitation of Sociomoral Reasoning in Delinquents. Journal of Consulting and Clinical Psychology, 52(1), 37–45.

Gibbs, J. C., Basinger, K., & Fuller, D. (1992). Moral Maturity: Measuring the Development of Sociomoral Reflection. New Jersey: Lawrence Erlbaum.

181

Gibbs, J. C., Basinger, K. S., Grime, R. L., & Snarey, J. R. (2007). Moral Judgment Development Across Cultures: Revisiting Kohlberg’s Universality Claims. Developmental Review, 27(4), 443–500.

Giddings, L. (2006). Mixed Methods Research: Positivism Dressed in Drag. Journal of Research in Nursing, 11(3), 195–203.

Glick B and Goldstein A. (1987). Aggression Replacement Training. Journal of Counseling and Development, 65, 356–362.

Glick, B., & Gibbs, J. (2011). Aggression Replacement Training (3rd ed.). Illinois: Research Press.

Goldstein, A., Glick, B., & Gibbs, J. (1998). Aggression Replacement Training (2nd ed.). Illinois: Research Press.

Goldstein, A., Glick, B., Reiner, S., Zimmerman, D., & Coultry, T. (1987). Aggression Replacement Training: A Comprehensive Intervention for Aggressive Youth. Illinois: Research Press.

Goldstein, A. P. (1973). Structured Learning Therapy. New York: Academic.

Goldstein, A. P. (1999). Low Level Aggression. Illinois: Research Press.

Goldstein, A. P., Palumbo, J., Striepling, S., & Voutisnas, A. M. (1995). Break It Up: A Teacher’s Guide to Managing Student Aggression. Illinois: Research Press.

Goldstein, A. P., Sprafkin, R. P., Gershaw, N. J., & Klein, P. (1980). Skillstreaming the Adolescent. Illinois: Research Press Company.

Gough, A., Oliver, S., & Thomas, J. (2012). An Introduction to Systematic Reviews. London: Sage.

Gough, D. (2007). Weight of Evidence: A Framework for the Appraisal of the Quality and Relevance of Evidence. Research Papers in Education, 22(2), 213–228.

Gray, D. (2005). Doing Research in the Real World (2nd Edition). London: Sage.

Gray, D. (2014). Doing Research in the Real World (3rd Edition). London: Sage.

Gray, P., Miller, A., & Noakes, J. (1994). Challenging Behaviour in Schools. London: Routledge.

Gregg, V., Gibbs, J., & Basinger, K. (1994). Patterns of Developmental Delay in Moral Judgment by Make and Female Delinquents. Merrill-Palmer Quarterly, 40(4), 538–553.

182

Greig, A. (2001). The Educational Psychologist as Practitioner-Researcher: Reality or Dream? Educational and Child Psychology, 18(4), 75–88.

Gresham, F., & Elliott, S. (1993). Social Skills Interventions Guide. Special Services in the Schools, 8(1), 137–158.

Gresham, F., & Elliott, S. (2008). SSIS Social Skills Improvement System. Minneapolis: Pearson Publications

Gresham, F. M., Cook, C. R., Crews, S. D., & Kern, L. (2004). Social Skills Training for Children and Youth with Emotional and Behavioral Disorders : Validity Considerations and Future Directions. Behavioral Disorders, 30(1), 32–46.

Guba, E., & Lincoln, Y. (1994). Competing Paradigms in Qualitative Research. In Denzin, N. And Lincoln, Y. (Eds) The Sage Handbook of Qualitative Research. California: Sage.

Guerra, N., & Bradshaw, C. (2008). Linking the Prevention of Problem Behaviors and positive youth development. In Guerra, N. & Bradshaw, C. (Eds.), Core competencies to prevent problem behaviors and promote positive youth development. New Directions for Child and Adolescent Development. San Francisco: Jossey-Bass.

Gundersen, K., Finne, J., & Olsen, T. (2006). ART a Method for Training Social Competence: Training Book. Rogaland: University College Rogaland.

Gundersen, K., Olsen, T. M., Finne, J., & Domben, E. (2014). Aggression Replacement Training: A Method for Developing Social Competence. Rogaland: Diakonhjemmet Hogskole.

Gundersen, K., & Svartdal, F. (2005). Evaluation of a Norwegian Postgraduate Training Programme for the Implementation of Aggression Replacement Training. Psychology, Crime & Law, 11(4), 435–444.

Gundersen, K., & Svartdal, F. (2006). Aggression Replacement Training in Norway: Outcome evaluation of 11 Norwegian student projects. Scandinavian Journal of Educational Research, 50(1), 63–81.

Gundersen, K., & Svartdal, F. (2010). Diffusion of treatment interventions: exploration of “secondary” treatment diffusion. Psychology, Crime & Law, 16(3), 233–249.

Gutman, L. M., & Vorhaus, J. (2012). The Impact of Pupil Behaviour and Wellbeing on Educational Outcomes (Vol. 2). Sheffield: DfE.

Hammersley, M. (2001). On “ Systematic ” Reviews of Research Literatures : A “ Narrative ” Response to Evans & Benefield. British Educational Research Journal, 27(5), 543–554.

183

Hargreaves, D. H. (1999). Revitalising Educational Research: Lessons from the Past and Proposals for the Future. Cambridge Journal of Education, 29(2), 239–249.

Harris, R. (1978). Relationship Between EEG Abnormality and Aggressive and Anti-Social Behaviour - A Critical Appraisal. In Aggression and Anti-Social Behaviour. In Hersov, L. and Burger, M. (Eds.) Childhood and Adolescence. Oxford: Pergamon Press.

Hatcher, R. M., Palmer, E. J., McGuire, J., Hounsome, J. C., Bilby, C. A. L., & Hollin, C. R. (2008). Aggression replacement training with adult male offenders within community settings: a reconviction analysis. Journal of Forensic Psychiatry & Psychology, 19(4), 517–532.

Hay, D. F., Payne, A., & Chadwick, A. (2004). Peer Relations in Childhood. Journal of Child Psychology and Psychiatry, and Allied Disciplines, 45(1), 84–108

Hersov, L. A., Berger, M., & Shaffer, D. (1978). Aggression and Anti-Social Behaviour in Childhood and Adolescence. Oxford: Pergamon Press.

HMSO. (2007). Aiming High for Young People: A Ten Year Strategy for Positive Activities. London: HMSO.

Hollin, C. R. (2004). Aggression Replacement Training: The Cognitive-Behavioural Context. In Goldstein, A. Nensen, R. Daleflod, B & Kalt, M (Eds.), New Perspectives on Aggression Replacement Training. West Sussex: John Wiley and Sons.

Holmqvist, R., Hill, T., & Lang, A. (2009). Effects of aggression replacement training in young offender institutions. International Journal of Offender Therapy and Comparative Criminology, 53(1), 74–92.

Home Office. (2013). Addressing Youth Violence and Gangs Practical Advice for Schools and Colleges. London: Home Office.

Hops, H. (1983). Children’s Social Competence and Skill: Current Research Practices and Future Directions. Behavior Therapy, 14, 3–18.

Huesmann, L. R., Dubow, E. F., & Boxer, P. (2009). Continuity of Aggression from Childhood to Early Adulthood as a Predictor of Life Outcomes: Implications for the Adolescent-Limited and Life-Course-Persistent Models. Aggressive Behavior, 35(2), 136–49.

Huesmann, L. R., Eron, L. D., Lefkowitz, M. M., & Walkder, L. O. (1984). Stability of Aggression Over Time and Generations. Developmental Psychology, 20(6), 1120–1134.

Hughes, J. (1988). Cognitive Behavior Therapy with Children in Schools. Oxford: Pergamon Press.

184

Indoe, D. (1995). Cognitive Behavioural Therapy and Children of the Code. Educational and Child Psychology, 12(4), 71–81.

Joffe, H. (2012). Thematic Analysis. In Harper, D & Thompson, A (Eds.), Qualitative Research Methods in Mental Health and Psychotherapy. West Sussex: John Wiley and Sons.

Johnson, D. W. & Johnson, F. P. (2009) Joining Together: Group Theory and Group Skills. New Jersey: Pearson Education.

Johnson, R. B., & Onwuegbuzie, A. J. (2004). Mixed Methods Research : A Research Paradigm Whose Time Has Come. Educational Researcher, 33(7), 14–26.

Jones, Y. (1991). Aggression Replacement in a High School Setting. Australian Journal of Guidance and Counselling, 1(1), 81–99.

Kam, C., Greenberg, M. T., & Walls, C. T. (2003). Examining the Role of Implementation Quality in School-Based Prevention Using the PATHS Curriculum. Prevention Science, 4(1), 55–63.

Kaunitz, C., Andershed, A. K., Brannstrom, L., & Smedslund, G. (in press). Aggression Replacement Training ( ART ) for Reducing Antisocial Behaviour in Adolescents and Adults. Retrieved June 13, 2014, from http://campbellcollaboration.org/lib/project/113/

Kazdin, A. E. (2003a) Clinical Psychology 4th Edition. Boston: Pearson

Kazdin, A. (2003b). Research Design in Clinical Psychology. Boston: Allyn & Bacon.

Kazdin, A. E. (1987). Treatment of Antisocial Behavior in Children : Current Status and Future Directions. Psychological Bulletin, 102(2), 187–203.

Kelly, B. (2008) Frameworks for Practice in Educational Psychology: Coherent Perspectives for a Developing Profession. In Kelly, B., Woolfson, L. & Boyle, J. (Eds.) Frameworks for Practice in Educational Psychology. London: Jessica Kingsley Publishers.

Kelly, B. (2012). Implementation Science for Psychology in Education. In Kelly, B & Perkins, D. (Eds.), Handbook of Implementation Science for Psychology in Education. USA: Cambridge University Press.

Kennedy, M. M. (1999). A Test of Some Common Contentions About Educational Research. American Educational Research Journal, 36(3), 511–541.

Kershaw, P., & Sonuga-Barke, E. (1998). Emotional and Behavioural Difficulties: Is This a Useful Category? Educational and Child Psychology, 15(4), 45–55.

185

Klinger, J., Ahwee, S., Pilonieta, P., & Menendez, R. (2003). Barriers and Facilitators in Scaling up Research-Based Practices. Exceptional Children, 69(4), 411–429.

Kohlberg, L. (1973). Collected Papers on Moral Development and Moral Education. United States of America: Harvard University.

Kohlberg, L. (1984). The Psychology of Moral Development. San Francisco: Harper & Row.

Kokko, K., & Pulkkinen, L. (2005). Stability of Aggressive Behavior from Childhood to Middle Age in Women and Men. Aggressive Behavior, 31(5), 485–497.

Koposov, R., Gundersen, K., & Svartdal, F. (2014). Efficacy of Aggression Replacement Training among Children from North-West Russia. The International Journal of Emotional Education, 6(1), 14–24.

Kratcoski, P. C. (1985). Youth Violence Directed Toward Significant Others. Journal of Adolescence, 8, 145–157.

Kratochwill, T. R. (2003). Task Force on Evidence Based Interventions in School Psychology. USA: APA.

Krishnan, P., See Yeo, L., & Cheng, Y. (2012). School-based Cognitive Behavioural Therapy for Academically Underachieving Singaporean Adolescents with Aggressive and Rule Breaking Behaviour. Asia Pacific Journal of Counselling and Psychotherapy, 4(1), 3–17.

Ladd, G. W. (1999). Peer Relationships and Social Competence During Early and Middle Childhood. Annual Review of Psychology, 50, 333–359.

Langeveld, J. H., Gundersen, K. K., & Svartdal, F. (2012). Social Competence as a Mediating Factor in Reduction of Behavioral Problems. Scandinavian Journal of Educational Research, 56(4), 381–399.

Leeman, L. W., Gibbs, J. C., & Fuller, D. (1993). Evaluation of a Multi-Component Group Treatment Program for Juvenile Delinquents. Aggressive Behavior, 19, 281–292.

Lewin, K. (1936). Principles of Topological Psychology. New York: McGraw-Hill.

Liau, A. K., Barriga, A. Q., & Gibbs, J. C. (1998). Relations Between Self-Serving Cognitive Distortions and Overt vs . Covert Antisocial Behavior in Adolescents. Aggressive Behavior, 24, 335–346.

Liddle, I., & Macmillan, S. (2010). Evaluating the FRIENDS Programme in a Scottish Setting. Educational Psychology in Practice, 26(1), 53–67.

186

Lochman, J. E. (2000). Theory and Empiricism in Intervention Research : A Dialectic to Be Avoided. Journal of School Psychology, 38(4), 359–368.

Lochman, J. E., & Dodge, K. a. (1994). Social-Cognitive Processes of Severely Violent, Moderately Aggressive and Nonaggressive Boys. Journal of Consulting and Clinical Psychology, 62(2), 366–374.

Lochman, J. E., & Wells, K. (2003). Effectiveness of the Coping Power Program and of Classroom Intervention With Aggressive Children " Outcomes at a 1-Year Follow-Up. Behavior Therapy, 34, 493–515.

Luria, A. (1961). The Role of Speech in the Regulation of Normal and Abnormal Behavior. New York: Liveright Publishing.

Malekoff, A. (2004). Group Work with Adolescents: Principles and Practice. New York: The Guilford Press.

Maras, P., & Kutnick, P. (1999). Emotional and Behavioural Difficulties in Schools : Consideration of Relationships Between Theory and Practice. Social Psychology of Education, 3, 135–153.

Marshall, S. (1996). The Characteristics of Cognitive Behaviour Therapy. In Marshall, S & Turnbull, J. (Eds.), Cognitive Behaviour Therapy. London: Bailliere-Tindall.

Marshall, S., & Turnbull, J. (1996). Cognitive Behaviour Therapy. London: Bailliere-Tindall.

McCrady, F., Kaufman, K., Vasey, M. W., Barriga, A. Q., Devlin, R. S., & Gibbs, J. C. (2008). It’s All About Me: A Brief Report of Incarcerated Adolescent Sex Offenders' Generic and Sex-Specific Cognitive Distortions. Sexual Abuse : A Journal of Research and Treatment, 20(3), 261–71.

Mcginnis, E. (2003). Aggression Replacement Training: A Viable Alternative. Reclaiming Children and Youth, 12(3), 161–166.

McGuire, J., & Clark, D. (2004). A National Dissemination Program. In Goldstein, A. Nensen, R. Daleflod, B & Kalt, M. (Eds.), New Perspectives on Aggression Replacement Training. West Sussex: John Wiley and Sons.

Meichenbaum, D. (1979). Cognitive-Behavior Modification. London: Plenum Press.

Meichenbaum, D., & Goodman, J. (1969a). Reflection-Impulsivity and Verbal Control of Motor Behavior (b). Child Development, 40, 785–797.

Meichenbaum, D., & Goodman, J. (1969b). The Developmental Control of Operant Motor Responding by Verbal Operants. Journal of Experimental Child Psychology, 7, 553–565.

187

Meichenbaum, D., & Goodman, J. (1971). Training Impulsive Children to Talk to Themselves: A Means of Developing Self-Control. Journal of Abnormal Psychology, 77(2), 115–126.

Mental Health Foundation. (1999). Bright Futures. Promoting Children and Young People’s Mental Health. London: The Mental Health Foundation.

Merrett, F., & Whedall, K. (1987). British Teachers and The Behavioural Approach. In Whedall, K. (Ed.), The Behaviourist in the Classroom. London: Allen and Unwin.

Mertens, D. (2010). Research and Evaluation in Education and Psychology. London: Sage.

Miller, A. (2003). Teachers, Parents and Classroom Behaviour. Berkshire: McGraw-Hill Education.

Morgan, D. L. (2007). Paradigms Lost and Pragmatism Regained: Methodological Implications of Combining Qualitative and Quantitative Methods. Journal of Mixed Methods Research, 1(1), 48–76.

Moynahan, L. (2003). Enhanced Aggression Replacement Training with Children and Youth with Autism Spectrum Disorder. Reclaiming Children and Youth, 12(3), 174–180.

Moynahan, L., & Stromgren, B. (2005). Preliminary Results of Aggression Replacement Training for Norwegian Youth with Aggressive Behaviour and with a Different Diagnosis. Psychology, Crime & Law, 11(4), 411–419.

National Offender Management Service. (2010). Young Offenders and the ART Programme. London: National Offender Management Service.

Nelson, D. A., & Crick, N. R. (1999). Rose-Colored Glasses: Examining the Social Information-Processing of Prosocial Young Adolescents. The Journal of Early Adolescence, 19(1), 17–38.

Novaco, R. (1975). Anger Control. Massachusetts: Lexington Books.

Novaco, R., Ramm, M., & Black, L. (2004). Anger Treatment with Offenders. In Hollin, C. (Ed.), The Essential Handbook of Offender Assessment and Treatment. West Sussex: John Wiley and Sons.

Novy, F., & McFarland, L. (2011). The Council on At-Risk Youth’s Aggression Replacement Training: Principles, Practices and Outcomes. 2009-2010 School Year. University of Texas: Inter-American Institute for Youth Justice.

Nugent, W. R., Bruley, C., & Allen, P. (1998). The Effects of Aggression Replacement Training on Antisocial Behaviour in a Runaway Shelter. Research on Social Work Practice, 8(6), 637–656.

188

Nugent, W. R., Bruley, C., & Allen, P. (1999). The Effects of Aggression Replacement Training on Male and Female Antisocial Behavior in a Runaway Shelter. Research on Social Work Practice, 9(4), 466–482.

Nugent, W. R., Champlin, D., & Wiinimaki, L. (1997). The Effects of Anger Control Training on Adolescent Antisocial Behaviour. Research on Social Work Practice1, 7(4), 446–462.

Ofsted. (2005). Managing challenging behaviour. London: Ofsted.

OfSted. (2013). Data Dashboard. Retrieved October 29, 2014, from http://dashboard.ofsted.gov.uk/

Pakaslahti, L., & Keltikangas-Jarvinen, L. (1996). Social Acceptance and the Relationship Between Aggressive Problem-Solving Strategies and Aggressive Behaviour in 14-year-old Adolescents. European Journal of Personality, 10(1996), 249–261.

Palmer, E. J., & Hollin, C. R. (1998). A Comparison of Patterns of Moral Development in Young Offenders and Non-Offenders. Legal and Criminological Psychology, 3, 225–235.

Parsons, C., & Castle, F. (1998). The Cost of School Exclusion in England. International Journal of Inclusive Education, 2(4), 277–294.

Pepler, D. J., & Rubin, K. H. (1991). The Development and Treatment of Childhood Aggression. New Jersey: Lawrence Erlbaum.

Pepler, D., King, G., & Byrd, W. (1991). A Social-Cognitively Based Social Skills Training Program for Aggressive Children. In Pepler, D. & Rubin, K. (Eds.), Development and Treatment of Childhood Aggression1. New Jersey: Lawrence Erlbaum.

Petticrew, M., & Roberts, H. (2006). Systematic Reviews in the Social Sciences - A Practical Guide. Oxford: Blackwell.

Phillips, A. A., & Chamberlain, V. (2006). MORI Five-Year Report : An analysis of Youth Survey Data. London: Youth Justice Board.

Piaget, J. (1965). The Moral Judgment of the Child. New York: The Free Press.

Prinz, R., Blechman, E., & Dumas, J. (1994). An Evaluation of Peer Coping-Skills Training for Childhood Aggression. Journal of Clinical Child psychology, 23(2), 193–203.

Pritchett, R., Pritchett, J., Marshall, E., Davidson, C., & Minnis, H. (2013). Reactive Attachment Disorder in the General Population : A Hidden ESSENCE Disorder. The Scientific World Journal, 2013, 1–12.

189

Prochaska, J. O., Norcross, J. C. & DiClemente, C. C. (1994) Changing for Good. The Revolutionary Program That Explains the Six Stages of Change and Teaches You How to Free Yourself From Bad Habits. New York: William Morrow and Company.

Quinn, M. M., Kavale, K. a., Mathur, S. R., Rutherford, R. B., & Forness, S. R. (1999). A Meta-Analysis of Social Skill Interventions for Students with Emotional or Behavioral Disorders. Journal of Emotional and Behavioral Disorders, 7(1), 54–64.

Rait, S., Monsen, J. J., & Squires, G. (2010). Cognitive Behaviour Therapies and Their Implications for Applied Educational Psychology Practice. Educational Psychology in Practice, 26(2), 105–122.

Razali, N. M., & Wah, Y. B. (2011). Power Comparisons of Shapiro-Wilk , Kolmogorov-Smirnov , Lilliefors and Anderson-Darling Tests, Journal of Statistical Modelling and Analytics. 2(1), 21–33.

Reid, S. (1987). The Use of Groups for Therapeutic Interventions. Educational and Child Psychology, 4(3/4), 171–179.

Renk, K., & Phares, V. (2004). Cross-informant Ratings of Social Competence in Children and Adolescents. Clinical Psychology Review, 24(2), 239–54.

Roberts, J. M. (2009). The Effectiveness of Aggression Replacement Training as Determined by the Attitudes Towards Guns and Violence Questionnaire. Unpublished Doctoral Dissertation, Alliant International University, California. United States of America.

Robson, C. (2011). Real World Research. West Sussex: John Wiley and Sons.

Rose-Krasnor, L., Rubin, K. H., Booth, C. L., & Coplan, R. (1996). The Relation of Maternal Directiveness and Child Attachment Security to Social Competence in Preschoolers. International Journal of Behavioral Development, 19(2), 309–325.

Ross, A., Duckworth, K., Smith, D. J., Wyness, G., & Schoon, I. (2011). Prevention and Reduction: A Review of Strategies for Intervening Early to Prevent or Reduce Youth Crime and Anti-Social Behaviour. London: DfE.

Rubin, K. H., Bream, L. A., & Rose-Krasnor, L. (1991). Social Problem Solving and Aggression in Childhood. In Pepler, D. & Rubin, K. (Eds.), The Development and Treatment of Childhood Aggression. New Jersey: Lawrence Erlbaum.

Sarason, I. G., & Sarason, B. R. (1981). Teaching Cognitive and Social Skills to High School Students. Journal of Consulting and Clinical Psychology, 49(6), 908–18.

190

Schrag, F. (1992). In Defense of Positivist Research Paradigms. Educational Researcher, 21(5), 5–8.

Scott, C., Shaw, M., & Joughin, A. (2001). Finding the Evidence: A Gateway to the Literature in Child and Adolescent Mental Health. London: Gaskell.

Semrud-Clikeman, M. (2007). Social Competence in Children. New York: Springer.

Shadish, W. R., Cook, T. D., & T, C. D. (2002). Experimental and Quasi Experimental Designs. USA: Wadsworth.

Sheehan, K., DiCara, J. a., LeBailly, S., & Christoffel, K. K. (1999). Adapting the Gang Model: Peer Mentoring for Violence Prevention. Pediatrics, 104(1), 50–54.

Slavin, R. (2002) Evidence-Based Education Policies: Transforming Educational Practice and Research. Educational Researcher, 37(1), 15-21.

Snarey, J. R. (1985). Cross-Cultural Universality of Social-Moral Development: A Critical Review of Kohlbergian Research. Psychological Bulletin, 97(2), 202–32.

Sofronoff, K., Attwood, T., & Hinton, S. (2005). A Randomised Controlled Trial of a CBT Intervention for Anxiety in Children with Asperger Syndrome. Journal of Child Psychology and Psychiatry, 46(11), 1152–60.

Solomon, R. W., & Wahler, R. G. (1973). Peer Reinforcement Control of Classroom Problem Behaviour. Journal of Applied Behavior Analysis, 6(1), 49–56.

Southam-Gerow, M. A., & Kendall, P. C. (2000). Cognitive-Behaviour Therapy with Youth : Advances , Challenges , and Future Directions. Clinical Psychology and Psychotherapy, 366, 343–366.

Squires, G. (2001). Using Cognitive Behavioural Psychology with Groups of Pupils to Improve Self-Control of Behaviour. Educational Psychology in Practice, 17(4), 317–335.

Stallard, P. (2007). Think Good-Feel Good. West Sussex: John Wiley and Sons.

Stams, G. J., Brugman, D., Deković, M., van Rosmalen, L., van der Laan, P., & Gibbs, J. C. (2006). The Moral Judgment of Juvenile Delinquents: A Meta-Analysis. Journal of Abnormal Child Psychology, 34(5), 697–713.

Stattin, H., & Magnusson, D. (1989). The Role of Early Aggressive Behavior in the Frequency, Seriousness, and Types of Later Crime. Journal of Consulting and Clinical Psychology, 57(6), 710–8. R

191

Stirling, M. (1991). How Many Pupils are Being Excluded? British Journal of Special Education, 19(4), 128–130.

Stoiber, K. C., & Kratochwill, T. R. (2000). Empirically Supported Interventions and School Psychology : Rationale and Methodological Issues — Part I. School Psychology Quarterly, 15(1), 75–105.

Stoiber, K., & Waas, G. (2002). A Contextual and Methodological Perspective on the Evidence-Based Intervention Movement Within School Psychology in the United States. Educational and Child Psychology, 19(3), 7–21.

Sudbeck, N. (2010). A Qualitative Study of Teacher Perspectives on the Usefulness, Effectiveness and Potential Limitations of Aggression Replacement Training at an Alternative Junior and Senior High School. Unpublished doctoral dissertation, Argosy University, Minnesota. United States of America.

Sukhodolsky, D. G., Kassinove, H., & Gorman, B. S. (2004). Cognitive-Behavioral Therapy for Anger in Children and Adolescents: a Meta-analysis. Aggression and Violent Behavior, 9(3), 247–269.

Tashakkori, A., & Teddlie, C. (2003). Handbook of Mixed Methods in Social and Behavioral Research. London: Sage.

Taylor, J. L., Novaco, R. W., Gillmer, B. T., Robertson, A., & Thorne, I. (2005). Individual Cognitive-Behavioural Anger Treatment for People with Mild-Borderline Intellectual Disabilities and Histories of Aggression: a Controlled Trial. The British Journal of Clinical Psychology, 44(3), 367–82.

Telford, A., & Farrington, A. (1996). Generalisation and Maintenance of Self-Management Skills. In Marshall, S. & Turnbull, J. (Eds.), Cognitive Behaviour Therapy. London: Bailliere-Tindall.

The Office of the Children’s Commissioner. (2012). “ They never give up on you .” London: The Office of the Children’s Commissioner.

Tierney, T., & Dowd, R. (2000). The Use of Social Skills Groups to Support Girls with Emotional Difficulties in Secondary Schools. Support for Learning, 15(2), 82–85.

Tomlinson, S. (2005). Education in a Post Welfare Society. Berkshire: McGraw-Hill Education.

UNICEF. (2007). Child Poverty in Perspective: An Overview of Child Well-Being in Rich Countries. Italy: UNICEF.

United States Department of Justice (n.d.). Model Programs. Retrieved March 03, 2014, from www.ojjdp.gov/mpg/Program

192

Vaughn, B. E., & Waters, E. (1981). Attention Structure, Sociometric Status, and Dominance: Interrelations, Behavioral Correlates, and Relationships to Social Competence. Developmental Psychology, 17(3), 275–288.

Westbrook, D., Kennerley, H., & Kirk, J. (2011). Cognitive Behaviour Therapy. London: Sage.

What Works Clearinghouse. (2011). Procedures and Standards Handbook. USA: IES.

Whitfield, G. W. (1999). Validating School Social Work: An Evaluation of a Cognitive-Behavioral Approach to Reduce School Violence. Research on Social Work Practice, 9(4), 399–426.

Zelli, A., Dodge, K. a, Lochman, J. E., & Laird, R. D. (1999). The Distinction Between Beliefs Legitimizing Aggression and Deviant Processing of Social Cues: Testing Measurement Validity and the Hypothesis that Biased Processing Mediates the Effects of Beliefs on Aggression. Journal of Personality and Social Psychology, 77(1), 150–66.

193

6.1 Secondary sources

Adler, P. A. & Adler, P. (1998) Peer Power: Preadolescent Culture and Identity.

In Lease, M.,Musgrove, K. T. and Axelrod. J. L. (2002) Dimensions of Social

Status in Preadolescent Peer Groups: Likability, Perceived Popularity and

Social Dominance. Social Development. 11(4). 508-533.

Colby, A. & Kohlberg, L. (1987) The Measurement of Moral Judgement:

Theoretical Foundations and Research Validation (Vol. 1) In Gibbs, J. C.,

Basinger, K. S. & Fuller D. (1992) Moral Maturity New Jersey: Lawrence

Earlbaum.

Sugg, D. (2000) Wiltshire Aggression Replacement Therapy (ART): One Year

Reconvictions and Targetting. In Maguire, J. and Clark, D. (2004) A National

Dissemination Programme. In Goldstein, A. P., Nensen, R., Daleflod, B. & Kalt,

M. New Perspectives on Aggression Replacement Training. West Sussex: John

Wiley.

Reynolds, C. R. & Kamphaus, R. W (2007) Behavior Assessment System for

Children-Second edition. In Gresham, F. M. & Elliott, S. N. (2008) Social Skills

Improvement System. Minneapolis: Pearson.

194

7. Appendices

7.1 Appendix I: Systematic Review: A Detailed Description of the

Process Undertaken

Fig. 7.1: A Flow Chart Depicting the Search Strategy Employed in the Systematic Review.

The search terms used were as follows:

For the four online databases:

‘Aggression Replacement Training’

‘Aggression Replacement Training’ and ‘evaluation’

The research question was devised.

Key search terms were created (see below) based on their relevance to the research question, with reference to the thesaurus

tools on the databases and the intervention handbook.

These terms were used to conduct the searches. 4 online databases and 1 search

engine (see below) were utilised in the search.

The titles and abstracts of these studies were initially screened to ensure the

intervention and sample were relevant to the research question. Papers were also

discarded if duplicated or not in English (the researchers home language)

Full text articles were obtained for the remaining papers and the

studies were filtered with reference to inclusion and exclusion criteria.

Total papers:

29

Total papers:

5

195

‘Aggression Replacement Training’ and ‘social competence’

‘Aggression Replacement Training’ and ‘social skills’

‘Aggression Replacement Training’ and ‘behaviour’

‘Aggression Replacement Training’ and ‘behavior’

Due to the volume of results gathered from general search engines fewer

searches were conducted when utilising Google Scholar:

‘Aggression Replacement Training’ in exact words used and

‘adolescents’ and ‘evaluation’ in the option ‘with all the words’

In order to ensure as comprehensive a search as possible, ‘Full text’ searches

or searches of ‘All fields’ were carried out on all of the databases utilised. The

date range was also set to post-1986 as the original programme was devised in

1987. Where possible searches included the criterion that the papers should be

written in English and published in a peer reviewed journal, in order to avoid

retrieving irrelevant papers. Table 7.1 displays the raw data gathered from each

of the databases/search engines.

Source Total papers retrieved

After initial screening

After inclusion/exclusion criteria applied

ASSIA (via Proquest) 34 6 2

Web of Science 72 5 1

PsycInfo (via OVID SP) 84 3 2

Wiley Online Library 360 0 0

Google Scholar 1060 15 0

TOTALS 1610 29 5 Table 7.1: Table to Show the Results of the Literature Search Conducted on 4

th and 5

th June

2014.

196

7.2: Appendix II- Descriptive map of the studies to aid the in-depth review of the research

STUDY SETTING, PARTICIPANTS

AND SELECTION PROCESS

DESIGN AND INTERVENTION

MEASURES OUTCOMES

Glick and Goldstein (1987) [USA]

Annsville Youth Centre. A residential facility for incarcerated youth 60 Males (all of the residents at the centre). Aged 14-17 years.

Cluster Randomised control trial 24 in ART condition 24 in ‘brief instruction’ comparison group 12 in no treatment control group Those in the 2 control groups then completed the intervention in a repeated measures design. 30 sessions over 10 weeks All staff oriented in ART. Some fully trained as implementers.

(Gathered from Goldstein et al. 1987) Pre and post-test with a follow up. Direct situations test which assesses the social skills taught in ART. Minimal generalisation situations test and extended generalisation situations test assess skills transfer. (all self-report, created by Goldstein et al 1987) Behaviour incident reports (staff report, created by Goldstein et al. 1987) Self-control scale (staff report) Moral reflection measure (self-report)

No change in moral reasoning ART group acquired and transferred the social skills at significantly greater levels than the 2 control groups. ART group also experienced lower impulsivity and less behaviour incidents, at lower intensity, compared to the control groups. Repeated measures analyses of the 36 controls also showed significant decreases in intensity and number of acting out incidents. When released those from the ART groups scored higher on a measure of community functioning compared to controls.

Jones (1991) [Australia]

Suburban high school in a low socio-economic area with high crime rates. 18 students from

Randomised Control Trial. 18 students were randomly allocated to 3 groups of 6: ART group Comparison group (moral reasoning only)

Pre and post-test Self-control scale and behaviour incident reports (Goldstein et al 1987) completed by staff. Moral reflection measure and situations tests of social skills acquisition (Goldstein et al 1987)

ART group showed greatest decrease in aggression compared to the other 2 groups on the behaviour incident reports Both the treatment and moral reasoning groups showed an increase in coping incidents

197

years 8 and 9. Selected by form tutors as being ‘significantly disruptive’. Then rated on an aggression scale. 9 males and 9 females with highest aggression selected.

Control group (no treatment) 30 sessions over 10 weeks

completed by the students on a one to one basis

Control and ART groups showed greatest improvement on the combined self-control and impulsivity measure. No significant changes in moral reasoning. ART group acquired 3 skills on the situations tests, control group acquired 2 skills.

Coleman, Pfeiffer and Oakland (1992) [USA]

Residential treatment centre for adolescents with behaviour problems. 39 participants aged 13-18 years. 10 females, 29 males. 2/3rds were diagnosed with conduct disorder. Staff selected the students based on them displaying aggressive behaviours and self-control problems. 25% attrition rate (originally 52 subjects).

Randomised control trial. Pupils were randomly assigned to groups of 6, each facilitated by 2 trained members of staff from the centre. After attrition, 24 remained in the ART groups and 15 in the no treatment control groups. 50 ART sessions were spread over 10 weeks. The extra 20 sessions were assigned to homework and practice of the skills learnt. Integrity checks included daily logs completed by the facilitators and 2 observations per group by the first author and an ART trainer.

Pre and post-data was collected on 4 measures: Self-report measures included situations tests (Goldstein et al 1987) to measure social skills and a moral reflection measure. These were completed on a one to one basis with research assistants who were blind to the research groups. Staff report measures included a self-control scale and a social skills checklist, as well as behaviour incident reports (adapted from Goldstein et al 1987). Predictor measures (carried out at pre-test only) included a measure of antisocial behaviour and a personality index of locus of control and self-perception.

Only 1 measure resulted in improvement for the ART group over the control group. This was the social skills knowledge from the direct situations test (Goldstein et al. 1987). Specifically the skills of expressing a compliant, keeping out of fights and responding to group pressure. Personality variables (particularly the self-concept scales) did predict post-test self-control.

Gundersen and Svartdal (2006) [Norway]

65 young people with behaviour problems aged 11-17 years (16

Cluster Randomised Control Trial Students screened with a checklist of behaviour problems and groups composed so that

Pre and post. Parents, teachers and children completed a measure of behavioural functioning and one of social skills.

Parent and teacher measures of social skills suggested a significant increase for the ART group, whilst no change was found for the control

198

females and 49 males) Schools and institutions (including special schools and psychiatric clinics).

members in each group had varying levels of need. These groups were then randomly assigned to the conditions. (5 of the settings had both an intervention and control group) 47 received ART and 18 acted as a no treatment control group. 11 groups of students, who were conducting ART as part of further education course acted as facilitators. 24 sessions of ART over 13 weeks. On average this entailed 10.8 sessions of social skills training, 8.4 sessions of anger control and 4.8 of moral reasoning Integrity checks included supervision and questionnaires for the leaders.

Parents and teachers also completed a scale of behaviour problems and attention skills. The young people also completed a measure of cognitive distortions and a self-report questionnaire designed by the researchers, to assess problem behaviours and pro-social skills

groups. Social skills on the custom self-report measure also suggested a significant increase for the ART group. Both groups improved on the cognitive distortions test Parent, teacher and child (custom-made measure) reports of behaviour problems showed a significant decrease for the ART members which was not reflected in the controls. Similar results were found in the behavioural functioning measure. However, the self-report found that both groups experienced a significant decrease. In terms of moderating factors. Only 2 scales were moderated by age (the teacher problem behaviours scale and the parent behavioural functioning scale). The authors suggested that these outcomes did not impact on the conclusions of the research. There was no effect of institution on the outcomes.

Currie et al (2009) [Australia]

Youth Justice Custodial Setting 5 males aged 17-18 (initially 6 but one left the group). Participants were referred onto the programme by health workers (i.e.

One group, repeated measures. ART was facilitated by the provisional psychologist (the primary researcher), trained in ART and a social worker colleague 10 week programme

Pre and post-test All self-report measures including a questionnaire regarding aggression, a measure of cognitive distortions and a measure of social skills. All administered in one to one interviews.

Significant decrease in aggression, increase in self-control and overall social skills rating but no changes found in cognitive distortions.

199

psychologists) All had committed violent offences

Table 7.2: Descriptive Map of the Studies Found in the Systematic Literature Review

200

7.3 Appendix III: Quality Assessment: A description of the Weight of

Evidence criteria employed to review the studies gathered from the

systematic literature search.

Quality assessment

The Weight of Evidence Model created by (Gough 2007) was adopted in order

to judge the quality and relevance of the studies reviewed, so as to ascertain

the relative value of the findings in contributing towards answering the research

question.

The model utilises three sets of judgements, which are then combined to form

an overall assessment rating.

These criteria were designed with reference to What Works Clearinghouse

(2011) ‘Evidence standards for RCTs and Comparison Group Quasi

Experimental Designs’ (p.11) and Krachtowill (2003) ‘Key Features for Coding

Studies and Rating Level of Evidence/Support’ (p.26).

In the following categories if a study possesses a characteristic in the ‘high’

rating it is worth 3 points, ‘medium’ is 2 points and ‘low’ is 1 point. These are

then averaged together to provide a rating of low, medium of high for each

‘weight of evidence’ judgement (A, B and C).

High- 2.5-3

Medium-1.5-2.4

Low-1-1.4

Weight of Evidence A-The Quality of the Methodology

High-Multiple integrity checks; clear explanation of intervention procedures

which adhered to the ‘best practice’ scheduling prescribed by the handbook,

with a minimum of 30 sessions over 10 weeks; pre, post and follow up

201

measures; multiple measures from multiple sources, all of which are valid and

reliable; low attrition rates (<20% at post and <30% at follow up).

Medium-An integrity check; details provided regarding intervention procedures

but they did not follow the prescribed 30 sessions in 10 weeks; pre and post-

measure; multiple measures may be from a single source or some may not be

validated or tested for reliability; may be high attrition (>20%).

Low-No mention of or modified intervention procedures followed; no mention of

integrity checks; post-measures only; single outcome measure; no mention of

attrition.

Weight of Evidence B- The Relevance of the Methodology for Answering the

Review Question

High- Randomised Control Trial design; pre, post and follow up measures;

active comparison group (possibly with a no treatment control group also).

Medium- Cluster Randomised Control Trial; pre and post-measures; no-

treatment control group only.

Low- Quasi-experimental design; post-measures; no control group.

Weight of Evidence C-The Relevance of the Evidence for Answering the

Review Question

High- sample aged 10-19; an Aggression Replacement Training intervention

conducted over 10 weeks, at least 3 sessions per week, with equal numbers of

sessions for each component; participants who had been identified as having

difficulties in the areas targeted by the intervention (e.g. anger control,

aggression, social skills, moral reasoning); quantitative measures of behaviour

problems (including aggression), social skills and moral reasoning; conducted

by a trained facilitator.

Medium-an Aggression Replacement Training intervention which may be

shorter than the prescribed 30 sessions; 2 quantitative measures of behaviour

problems (including aggression), social skills or moral reasoning; little selection

to target individuals who need such support.

202

Low-sample aged below 10 and above 19; an adapted Aggression

Replacement Training intervention or no mention of intervention procedures; a

single quantitative measures of behaviour problems (including aggression),

social skills or moral reasoning; provided in a preventative manner (no selection

to identify the needs of the participants); conducted by an untrained facilitator or

no information regarding who facilitated.

Weight of Evidence D- Overall Assessment Rating.

In the above categories a ‘high’ rating is worth 3 points, ‘medium’ is 2 points and

‘low’ is 1 point.

This category ranks the studies in terms of their average score on the other 3

judgements.

High- 2.5-3

Medium-1.5-2.4

Low-1-1.4

203

7.4: Appendix IV Recruitment Leaflet and Initial Application Form from

the Educational Psychology Service to Support the Selection of Schools

onto the Project.

204

205

206

7.5 Appendix V: Information Leaflet for Schools Regarding the

Evaluation Research Project.

Information sheet for Schools

University of Nottingham School of Psychology

Research Project on a Multi component, Social Competence Intervention based on Cognitive Behaviour Therapy

Researcher: Samantha Grimes (Trainee Educational Psychologist) Email: Samantha.Grimes@birmingham.gov.uk

Telephone: 0121 303 8288. University supervisor: Neil Ryrie. neil.ryrie@nottingham.ac.uk

The purpose of the current research study is to investigate the effectiveness of

a Social Competence Training intervention (based on Aggression Replacement

Training®) in improving the pro-social skills and problem behaviours displayed

by pupils in school and their moral reasoning ability.

With support from the researcher each school will identify 6 students aged

between 11 and 18 years to take part in the intervention. These pupils should

be young people who consistently display some deficiencies in pro-social skills,

anger control and moral reasoning capacities. The students must also have

good attendance, not be involved in any other behaviour interventions and have

the ability to reflect upon their thoughts and behaviour. In order to identify 3

‘target’ individuals and 3 ‘role models’, who have slightly more advanced levels

of social competence, the school will be asked to select 3 pupils with ‘low level’

social behaviour difficulties and 3 pupils with more challenging behaviour

needs.

As part of the research the pupils will be asked to complete measures on two

occasions as a group: once in September prior to the intervention and once in

December following the intervention. These measures will be taken at the same

time whether the pupils are receiving the intervention in the Spring or Autumn.

207

These sessions require a small, quiet room and should take no more than 1

hour. Support from a Teaching Assistant would also be appreciated to support

those with literacy difficulties. On these two occasions a member of staff, who

has regular contact with the child, will also be asked to complete a

questionnaire about the pupil’s recent behavior in school. Checklists will also be

sent home to parents. These should take no more than 15-20 minutes. Finally

at the end of the project some students from the Autumn intervention groups

may be selected to take part in a short (20 minute) interview to gather their

views about the intervention.

If you have any questions or concerns please do not hesitate to ask. I can also

be contacted after your participation using the details listed above. Many thanks

for your time.

208

Aggression Replacement Training Research

Timeline-Autumn Schools

Selection process: 6 pupils, aged 11-

18 years. Half will display challenging

deficiencies in the prosocial skill,

anger control and moral reasoning

capacities and half experience more

‘low level’ difficulties.

MUST have good attendance, be able

to reflect on their thoughts and

behaviour and not be involved in

other interventions.

July

Once parental consent to take part in

the intervention from the EPS has

been obtained send parental

consent forms for the research

Gather all back in before summer

holidays

Once parental consent for the

research is obtained gain consent

from the pupils in the groups and

teachers who will complete the

teacher measures

Before the first session of measures

which will take place during the 3rd

week of September. Preferably

before the summer holidays.

Room and TA support for 1 hour for

the 2 sessions of measures

During the 3rd week of September

and final 2 weeks before the

Christmas holidays

Distributing and collecting the parent

and teacher questionnaires

During the 3rd week of September

and final 2 weeks before the

Christmas holidays

Completing an information sheet with At the same time as the first session

209

Sam about the participants (i.e. age,

academic levels, rate of exclusion

etc..)

of measures (3rd week Sept)

During the intervention period Sam

will visit 3 sessions to carry out

‘integrity checks’

September-December

Following the intervention, a pupil

interview will be conducted with half

of the group members (3) in order to

gain their views of the programme. I

would be grateful for the use of a

room for an hour and support with

gaining parent/pupil consent for this

interview

January

210

Aggression Replacement Training Research

Timeline-Spring Schools

Selection process: 6 pupils, aged 11-

18 years. Half will display challenging

deficiencies in the prosocial skill, anger

control and moral reasoning capacities

and half experience more ‘low level’

difficulties.

MUST have good attendance, be able

to reflect on their thoughts and

behaviour and not be involved in other

interventions.

July

Once parental consent to take part in

the intervention from the EPS has

been obtained send parental consent

forms for the research

Gather all back in before summer

holidays

Once parental consent for the

research is obtained gain consent

from the pupils in the groups and

teachers who will complete the

teacher measures

Before the first session of measures

which will take place during the 3rd

week of September. Preferably before

the summer holidays

Room and TA support for 1 hour for

the 2 sessions of measures

During the 3rd week of September and

final 2 weeks before the Christmas

holidays

Distributing and collecting the parent

and teacher questionnaires

During the 3rd week of September and

final 2 weeks before the Christmas

holidays

Completing an information sheet with At the same time as the first session of

211

Sam about the participants (i.e. age,

academic levels, rate of exclusion

etc..)

measures (3rd week Sept)

212

7.6 Appendix VI: Information Letter and Consent Form for Parents

Information sheet for parents

University of Nottingham

School of Psychology

Research Project on a Multi Component Social Competence Intervention based on Cognitive Behaviour Therapy

Researcher: Samantha Grimes (Trainee Educational Psychologist)

Email: Samantha.Grimes@birmingham.gov.uk

Telephone: 0121 303 8288.

University supervisor: Neil Ryrie. Neil.ryrie@nottingham.ac.uk

Dear Parent/Guardian

As part of a new initiative Birmingham Educational Psychology Service are

trialing a social competence intervention (based on Aggression Replacement

Training®) in a small sample of schools around the City. I believe that your child

has been identified by the school as someone who may benefit from the

programme and your consent has been requested for them to take part in the

intervention.

As part of this new intervention project we are running a piece of research to

determine its effectiveness. This research is supported by the Birmingham

Educational Psychology Service and The University of Nottingham, which I

attend as a Trainee Educational Psychologist. The aim is to ascertain how

effective the intervention is at improving the pro-social skills and problem

behaviours displayed by young people and their moral reasoning ability.

Your child will have the opportunity to consent to take part in the research

themselves and ask any questions that they might have nearer the start of the

programme. They will be asked to complete 2 questionnaires on two occasions

across the course of the project, once in September and once in December,

whether they are receiving the intervention in the Autumn or Spring, to compare

213

their results. This process should take approximately 45 minutes and will be

carried out during school time.

The children will be asked to complete two questionnaires during each of these

sessions. The first measure asks pupils to rate themselves on several aspects

of social skills, such as communication and cooperation as well as problem

behaviours, such as failing to control their temper, impulsive reactions and

feeling anxious. Examples of questions include: ‘I stay calm when dealing with

problems’, ‘I have temper tantrums’ and ‘I feel sad’. The second measure asks

pupils to consider some moral dilemmas such as: ‘Let’s say a friend of yours

needs help and may even die, and you’re the only person who can save him or

her. How important is it for a person (without losing his or her own life) to save

the life of a friend?’. The pupils respond by indicating how important they feel

this act is and give a reason why.

I am also asking teachers to complete short questionnaires at these times,

similar to the first measure that the pupil’s will complete, with the addition of

questions regarding the pupil’s academic competence. Additional data such as

rates of exclusion, academic achievement, date of birth, free school meal

eligibility and attendance at the sessions will also be gathered to aid the

analysis of the intervention measures.

In order to enhance the conclusions drawn from the research we are gathering

parental measures as part of the study and would be grateful if you could

complete a 20 minute questionnaire surrounding your child’s behaviour, similar

to the first measure which the children will be completing. This includes

questions about your child’s social skills and problem behaviours, such as

‘follows your directions’, ‘acts without thinking’ and ‘says bad things about self’.

This questionnaire will be provided to you by the school during the third week of

September and second week of December. I would be grateful if you could

complete and return these questionnaires within a week from when they are

handed out.

Throughout the project any information gathered regarding your child will be

kept confidential. All information written in any reports will ensure anonymity so

214

that they cannot be identified. Should the questionnaires raise any concerns,

the school’s SENCo will be contacted in order to arrange support for your child,

if they consider this appropriate. If, after signing the consent, you or your child

wish to withdraw from the study then you may do so at any time and all

information gathered up to this point will be destroyed. Summaries of the main

findings will be made available at the end of the project.

To make sure that your child has the opportunity to contribute to the research

please complete the consent form attached to this letter as soon as possible (no

later than 15th July) and return it to the school office. If you have any questions

or concerns please do not hesitate to enquire at your child’s school or contact

myself using the details listed at the top of this letter. Many thanks for your

support.

Yours sincerely,

Samantha Grimes

Trainee Educational Psychologist, Birmingham Educational Psychology

Service.

215

Parent Consent Form

Research Project on a Multi Component Social Competence Intervention based on Cognitive Behaviour Therapy

Researcher: Samantha Grimes (Trainee Educational Psychologist) Email: Samantha.Grimes@birmingham.gov.uk

Telephone: 0121 303 8288. University supervisor: Neil Ryrie. neil.ryrie@nottingham.ac.uk

Childs name: ________________________________

Year: ______________________________________

Please cross out as necessary:

Have you read and understood the information sheet YES/NO

Have you had the opportunity to ask questions and discuss the study YES/NO

Have all the questions been answered satisfactorily YES/NO

Have you received enough information about the study YES/NO

Do you understand that you and your child are free to withdraw from the study at any time without having to give a reason YES/NO

Do you agree to take part in the study YES/NO

Do you give permission for your child to take part in the study (subject to them giving consent also) YES/NO

‘This study has been explained to me to my satisfaction, I agree to take part and

give permission for my child to participate. I understand that myself and my

child are free to withdraw at any time.’

Signed: ____________________________________

Print name: _________________________________

Date: ______________________________________

216

7.7 Appendix VII: Information Letter and Consent Form for Teaching

Staff

Information sheet for Teachers

University of Nottingham

School of Psychology

Research Project on a Multi Component, Social Competence Intervention based on Cognitive Behaviour Therapy

Researcher: Samantha Grimes (Trainee Educational Psychologist)

Email: Samantha.Grimes@birmingham.gov.uk

Telephone: 0121 303 8288.

University supervisor: Neil Ryrie. neil.ryrie@nottingham.ac.uk

The purpose of the current research study is to investigate the effectiveness of

a Social Competence Training intervention (based on Aggression Replacement

Training®) in improving the pro-social skills and problem behaviours displayed

by pupils in school and their moral reasoning ability.

Each school will identify 6 students aged between 11 and 18 years to take part

in the intervention. All of these pupils will be young people who consistently

display some deficiencies in pro-social skills, anger control and moral reasoning

capacities. However, 3 members of the group will be acting as ‘positive role

models’ and therefore may have greater social competence than the other

pupils.

The programme itself is a 10 week curriculum, with 3x1 hour session per week.

Further information can be obtained from

http://www.aggressionreplacementtraining.org. It is intended that this will be

implemented in some schools between September and December 2014 and

between January and March 2015 in others.

As a member of staff who has regular contact with one (or more) of the young

people participating in the study, I would be grateful if you could complete a

217

short questionnaire about the pupil’s recent behavior in school on two

occasions: Once in September and once in December. This should take no

more than 15-20 minutes. If you agree to take part please could you complete

the attached consent form and return to ……………………. No later than 12th

September.

If you have any questions or concerns please do not hesitate to ask. I can also

be contacted after your participation using the details listed above. Many thanks

for your time, Samantha

218

Teacher Consent Form

Research Project on a Multi Component Social Competence Intervention based on Cognitive Behaviour Therapy

Researcher: Samantha Grimes (Trainee Educational Psychologist) Email: Samantha.Grimes@birmingham.gov.uk

Telephone: 0121 303 8288. University supervisor: Neil Ryrie. neil.ryrie@nottingham.ac.uk

Please cross out as necessary:

Have you read and understood the participant information sheet YES/NO

Have you had the opportunity to ask questions and discuss the study YES/NO

Have all the questions been answered satisfactorily YES/NO

Have you received enough information about the study YES/NO

Do you understand that you are free to withdraw from the study at any time, without having to give a reason: YES/NO

Do you agree to take part in the study YES/NO

‘This study has been explained to me to my satisfaction, and I agree to take

part. I understand that I am free to withdraw at any time.’

Signature of the Participant: Date:

Name (in block capitals)

219

7.8 Appendix VIII: Information Letter and Consent Form for

Participants in the Experimental Group

Pupil information sheet

NAME:

DATE OF BIRTH:

SCHOOL:

CLASS:

Your parent/carer received a letter last week explaining that a new project

would be running in school during the next academic year and they sent a letter

back saying that they would like you to join in. The school also thought that you

would benefit from taking part in the group.

The project means that you will be working with other children from your school,

taking part in some fun activities three times a week for 10 weeks. These

sessions will give you lots of different ways to calm down when you are feeling

angry, deal with difficult situations and maintain friendships.

Sam’s part in the project is to measure whether the groups were useful. I would

like to see if the games and activities work well and whether you thought the

group was helpful for you because we might want to do this in other schools. To

find this out I will be asking you to answer some questions at two different

times, once in September and after the sessions in December. Your teachers

and parents are answering some questions for me too.

At the end of the project I will write a report about the group and how well it

worked. When I write this report I will leave your name out so that no one can

work out who you are and I will keep all of your information locked away

somewhere safe.

220

If you decide during these 2 sessions with Sam that you do not want to answer

the questions anymore you can change your mind whenever you like, just tell

your teachers or parents or me.

221

Pupil consent form

Name: ________________________________

Year: ______________________________________

Please cross out as necessary:

Have you read and understood the information sheet YES/NO

Have you had the opportunity to ask questions and discuss the study YES/NO

Have all the questions been answered so that you understand what you are being asked to do YES/NO

Have you received enough information about the study YES/NO

Do you understand that you are free to leave the question sessions at any time without having to give a reason YES/NO

Do you agree to take part in the study YES/NO

‘This study has been explained to me and I understand what I will be doing. I

agree to take part in Sam’s questions. I understand that I can choose not to

answer these questions at any time.’

Signed: ____________________________________

Print name: _________________________________

Date: ______________________________________

Helped by: __________________________________

222

7.9 Appendix IX: Information Letter and Consent Form for Participants

in the Control Group

Pupil information sheet-controls

NAME:

DATE OF BIRTH:

SCHOOL:

CLASS:

Your parent/carer received a letter last week explaining that a new project

would be running in school during the next academic year and they sent a letter

back saying that they would like you to join in. The school also thought that you

would benefit from taking part in the group.

The project means that you will be working with other children from your school,

taking part in some fun activities three times a week for 10 weeks. These

sessions will give you lots of different ways to calm down when you are feeling

angry, deal with difficult situations and maintain friendships.

Sam’s part in the project is to measure whether it was useful. In order to do this,

before you take part in the groups in January I will be coming to your school on

two separate occasions to ask you to answer some questions. Once in

September and once in December. Your teachers and parents are answering

some questions for me too.

At the end of the project I will write a report about the information I gather.

When I write this report I will leave your name out so that no one can work out

who you are and I will keep all of your information locked away somewhere

safe.

223

If you decide during the 2 sessions with Sam that you do not want to answer the

questions anymore you can change your mind whenever you like, just tell your

teachers or parents or me.

224

Pupil consent form

Name: ________________________________

Year: ______________________________________

Please cross out as necessary:

Have you read and understood the information sheet YES/NO

Have you had the opportunity to ask questions and discuss the study YES/NO

Have all the questions been answered so that you understand what you are being asked to do YES/NO

Have you received enough information about the study YES/NO

Do you understand that you are free to leave the question sessions at any time without having to give a reason YES/NO

Do you agree to take part in the study YES/NO

‘This study has been explained to me and I understand what I will be doing. I

agree to take part in Sam’s questions. I understand that I can choose not to

answer these questions at any time.’

Signed: ____________________________________

Print name: _________________________________

Date: ______________________________________

Helped by: __________________________________

225

7.10 Appendix X: Information Letter and Consent Form for Educational

Psychologists Facilitating the Groups in the Experimental Condition.

Information sheet for Educational Psychologists

University of Nottingham School of Psychology

Research Project on a Multi component, Social Competence Intervention based on Cognitive Behaviour Therapy

Researcher: Samantha Grimes (Trainee Educational Psychologist) Email: Samantha.Grimes@birmingham.gov.uk

Telephone: 0121 303 8288. University supervisor: Neil Ryrie. neil.ryrie@nottingham.ac.uk

The purpose of the current research study is to investigate the effectiveness of

an Aggression Replacement Training intervention in improving the pro-social

skills and problem behaviours displayed by pupils in school and their moral

reasoning ability.

With support from the researcher each school will identify 6 students aged

between 11 and 18 years to take part in the intervention. These pupils should

be young people who consistently display some deficiencies in pro-social skills,

anger control and moral reasoning capacities. The students must also have

good attendance, not be involved in any other behaviour interventions and have

the ability to reflect upon their thoughts and behaviour. In order to identify 3

‘target’ individuals and 3 ‘role models’, who have slightly more advanced levels

of social competence, the school will be asked to select 3 pupils with ‘low level’

social behaviour difficulties and 3 pupils with more challenging behaviour

needs.

As part of the research the pupils will be asked to complete measures on two

occasions as a group: once in the 3rd week of September prior to the

intervention and once in December after the intervention has finished. On these

two occasions a member of school staff, who has regular contact with the child,

226

will also be asked to complete a questionnaire about the pupil’s recent behavior

in school. Checklists will also be sent home to parents. Finally at the end of the

project some students may be selected to take part in a short interview to

gather their views about the intervention.

During the Autumn sessions 3 integrity checks will be completed by the

researcher, one for each ‘type’ of session (skillstreaming, anger control and

moral reasoning), to ensure the validity of the research. As part of the sessions

please could those conducting the ‘Autumn’ intervention keep a measure of

attendance for the students, so that the impact of attendance on outcomes can

be analysed.

I would be grateful if, at the end of the programme, you would complete a short

questionnaire designed to elicit your views about the intervention, which will be

used to support the interpretation the findings gathered from the measures

outlined above. This should take no more than 10 minutes. If you are willing to

partake in this research please could you complete the consent form attached

and return it to Samantha Grimes, Trainee EP, based at the Oakhill Centre by

12th July.

If you have any questions or concerns please do not hesitate to contact me

using the details at the top of this letter. I can also be contacted after your

participation using these details. Many thanks for your time, Sam.

227

Educational Psychologist Consent Form.

Research Project on a Multi Component Social Competence Intervention based on Cognitive Behaviour Therapy

Researcher: Samantha Grimes (Trainee Educational Psychologist) Email: Samantha.Grimes@birmingham.gov.uk

Telephone: 0121 303 8288. University supervisor: Neil Ryrie. neil.ryrie@nottingham.ac.uk

Please cross out as necessary

• Have you read and understood the participant information sheet YES/NO

• Have you had the opportunity to ask questions and discuss the study YES/NO

• Have all the questions been answered satisfactorily YES/NO

• Have you received enough information about the study YES/NO

• Do you understand that you are free to withdraw from the study at any time

without having to give a reason YES/NO

• Do you agree to take part in the study YES/NO

‘This study has been explained to me to my satisfaction, and I agree to take

part. I understand that I am free to withdraw at any time.’

Signature of the Participant:

Date:

Name (in block capitals):

228

7.11 Appendix XI: Summary of the Structure of the ART Sessions

(summarised from Gundersen et al. 2014)

External Structure (common to all three components)

At the start:

Welcome to the group members with a reminder of rules

Review of the previous session

Review homework

At the end:

ART game

Review of the session

Wind up: evaluation of group performance or a ‘friendship round’ where

group members praise one another on their participation in the session;

briefing on the next session; handing out rewards; ART cheer

Internal Structure (specific to each component)

Anger Control Sessions:

Define the day’s skill

Facilitators demonstrate the skill

The need for the skill is discussed

Main players and co-players of the role play are selected

Role plays are planned and observation tasks are delegated to the

remaining members of the group

229

Role plays are conducted- ‘bubble talk’, where the actors pause to

comment on their thoughts and feelings aloud to the observers, is used

to make the actors cognitions visible during the role play

Feedback is gathered from the observers and from members of each

team of the role play

The next teams plan and conduct their role plays and receive feedback

until every young person has had the opportunity to be a ‘main player’ in

a role play

Homework is distributed

Prosocial Skills Sessions:

Same steps as the anger control sessions.

Moral Reasoning Sessions:

Introduce the vignette in which a moral dilemma is experienced

Cultivate mature morality: those with mature responses are asked to

state their reasoning first.

Challenge or remediate moral developmental delay: Those who respond

immaturely are asked why they responded differently to some questions.

Those with mature responses are invited to share their oopinions of

other’s views

Facilitators pose new questions to the group, challenging errors in

thinking and ask the group to reconsider, cultivating mature morality and

challenging developmental delay

Reinforce mature moral cognition by leading the group to agree on

principles and statements i.e. ‘can we agree that...’

Homework

230

7.12 Appendix XII: Integrity Checklists from the international Center

for Aggression Replacement Training (iCART)

Retrieved November 2013 from

www.aggressionreplacementtraining.org/HOME

231

232

233

234

235

236

237

7.13 Appendix XIII: The Socio-Moral Reflection Measure (Gibbs,

Basinger & Fuller, 1992)

Social Reflection Questionnaire

PP No. __________________ ___ Date:____________________

School name: _______________________

INSTRUCTIONS

In this questionnaire, we want to find out about the things you think are important for

people to do, and especially why you think these things (like keeping a promise) are

important. Please try to help us understand your thinking by WRITING AS MUCH AS

YOU CAN TO EXPLAIN-EVEN IF YOU HAVE TO WRITE OUT YOUR EXPLANATIONS MORE

THAN ONCE. Don’t just write ‘same as before’. If you can explain better or use

different words to show what you mean, that helps us even more. Please answer all

the questions, especially the ‘why’ questions. If you need to, feel free to use the space

in the margins to finish writing your answers.

1. Think about when you’ve made a promise to a friend of yours. How important

is it for people to keep promises, if they can, to a friend?

Circle one: very important important not important

WHY IS THAT VERY IMPORTANT/IMPORTANT/NOT IMPORTANT (WHICHEVER ONE

YOU CIRCLED)?

--------------------------------------------------------------------------------------------------------------

--------------------------------------------------------------------------------------------------------------

--------------------------------------------------------------------------------------------------------------

--------------------------------------------------------------------------------------------------------------

---------------------------

2. What about keeping a promise to anyone? How important is it for people to

keep promises, if they can, even to someone they hardly know?

Circle one: very important important not important

WHY IS THAT VERY IMPORTANT/IMPORTANT/NOT IMPORTANT (WHICHEVER ONE

YOU CIRCLED)?

238

--------------------------------------------------------------------------------------------------------------

--------------------------------------------------------------------------------------------------------------

--------------------------------------------------------------------------------------------------------------

--------------------------------------------------------------------------------------------------------------

---------------------------

3. How about keeping a promise to a child? How important is it for parents to

keep promises, if they can, to their children?

Circle one: very important important not important

WHY IS THAT VERY IMPORTANT/IMPORTANT/NOT IMPORTANT (WHICHEVER ONE

YOU CIRCLED)?

--------------------------------------------------------------------------------------------------------------

--------------------------------------------------------------------------------------------------------------

--------------------------------------------------------------------------------------------------------------

--------------------------------------------------------------------------------------------------------------

---------------------------

4. In general, how important is it for people to tell the truth?

Circle one: very important important not important

WHY IS THAT VERY IMPORTANT/IMPORTANT/NOT IMPORTANT (WHICHEVER ONE

YOU CIRCLED)?

--------------------------------------------------------------------------------------------------------------

--------------------------------------------------------------------------------------------------------------

--------------------------------------------------------------------------------------------------------------

--------------------------------------------------------------------------------------------------------------

---------------------------

5. Think about when you’ve helped your mother or father. How important is it for

children to help their parents?

Circle one: very important important not important

WHY IS THAT VERY IMPORTANT/IMPORTANT/NOT IMPORTANT (WHICHEVER ONE

YOU CIRCLED)?

--------------------------------------------------------------------------------------------------------------

--------------------------------------------------------------------------------------------------------------

--------------------------------------------------------------------------------------------------------------

--------------------------------------------------------------------------------------------------------------

---------------------------

239

6. Let’s say a friend of yours needs help and may even die, and you’re the only

person who can save him or her. How important is it for a person (without

losing his or her own life) to save the life of a friend?

Circle one: very important important not important

WHY IS THAT VERY IMPORTANT/IMPORTANT/NOT IMPORTANT (WHICHEVER ONE

YOU CIRCLED)?

--------------------------------------------------------------------------------------------------------------

--------------------------------------------------------------------------------------------------------------

--------------------------------------------------------------------------------------------------------------

--------------------------------------------------------------------------------------------------------------

---------------------------

7. What about saving the life of anyone? How important is it for a person (without

losing his or her own life) to save the life of a stranger?

Circle one: very important important not important

WHY IS THAT VERY IMPORTANT/IMPORTANT/NOT IMPORTANT (WHICHEVER ONE

YOU CIRCLED)?

--------------------------------------------------------------------------------------------------------------

--------------------------------------------------------------------------------------------------------------

--------------------------------------------------------------------------------------------------------------

--------------------------------------------------------------------------------------------------------------

---------------------------

8. How important is it for a person to live even if that person doesn’t want to?

Circle one: very important important not important

WHY IS THAT VERY IMPORTANT/IMPORTANT/NOT IMPORTANT (WHICHEVER ONE

YOU CIRCLED)?

--------------------------------------------------------------------------------------------------------------

--------------------------------------------------------------------------------------------------------------

--------------------------------------------------------------------------------------------------------------

--------------------------------------------------------------------------------------------------------------

---------------------------

9. How important is it for people not to take things that belong to other people?

Circle one: very important important not important

WHY IS THAT VERY IMPORTANT/IMPORTANT/NOT IMPORTANT (WHICHEVER ONE

YOU CIRCLED)?

240

--------------------------------------------------------------------------------------------------------------

--------------------------------------------------------------------------------------------------------------

--------------------------------------------------------------------------------------------------------------

--------------------------------------------------------------------------------------------------------------

---------------------------

10. How important is it for people to obey the law?

Circle one: very important important not important

WHY IS THAT VERY IMPORTANT/IMPORTANT/NOT IMPORTANT (WHICHEVER ONE

YOU CIRCLED)?

--------------------------------------------------------------------------------------------------------------

--------------------------------------------------------------------------------------------------------------

--------------------------------------------------------------------------------------------------------------

--------------------------------------------------------------------------------------------------------------

---------------------------

11. How important is it for judges to send people who break the law to jail?

Circle one: very important important not important

WHY IS THAT VERY IMPORTANT/IMPORTANT/NOT IMPORTANT (WHICHEVER ONE

YOU CIRCLED)?

--------------------------------------------------------------------------------------------------------------

--------------------------------------------------------------------------------------------------------------

--------------------------------------------------------------------------------------------------------------

--------------------------------------------------------------------------------------------------------------

---------------------------

241

7.14 Appendix XIV: The Prompt Sheet Used in the Semi-Structured

Interviews with the Participants in the Experimental Groups.

Aggression Replacement Training

Semi-structured interview-pupil prompts.

Effectiveness; Do you think the ART group helped you? How?

Have you used what you learnt?

Which part have you used the most?

What parts do you think were effective?

Were there things that didn’t work for you?

Implementation; What did you think of the ART sessions?

How did you feel taking part in the groups?

What did you like about the sessions/not like?

What could be changed to make ART better?

242

7.15 Appendix XV: Questionnaire Distributed to the Educational

Psychologists Facilitating the Experimental Groups.

Aggression Replacement Training EP Questionnaire

Please could you complete the following questionnaire, providing as much detail as possible,

regarding your experience of implementing the ART programme. Group (school) name: Do you think the ART intervention was effective? What, in your opinion, contributed to these outcomes? (What worked well or did not work so well?) If you were to run this programme again what changes would you make? Any other comments?

243

7.16 Appendix XVI: Letter Confirming Ethical Approval Received from

the Ethical Committee of the University of Nottingham.

Ref: 499R

Wednesday, 09 July 2014

Dear Samantha Grimes & Neil Ryrie,

Ethics Committee Review

Thank you for submitting an account of your proposed research ‘An

Evaluation of Aggression Replacement Training: The impact of a

multi-component, CBT-based intervention on the problem

behaviours, pro-social skills and moral development of pupils in

English secondary schools’.

That research has now been reviewed by the Ethics Committee and I

am pleased to tell you that your submission has met with the

committee’s approval.

Final responsibility for ethical conduct of your research rests with

you or your supervisor. The Codes of Practice setting out these

responsibilities have been published by the British Psychological

Society and the University Research Ethics Committee. If you have

any concerns whatever during the conduct of your research then

you should consult those Codes of Practice.

Independently of the Ethics Committee procedures, supervisors also

have responsibilities for the risk assessment of projects as detailed

in the safety pages of the University web site. Ethics Committee

approval does not alter, replace, or remove those responsibilities,

nor does it certify that they have been met.

Yours sincerely

Dr Alan Sunderland

Chair, Ethics Committee

244

7.17 Appendix XVII Parental Information and Consent Form for the

Qualitative Measures

Parent Consent Form for Pupil Interviews

Research Project on a Multi Component Social Competence Intervention based on Cognitive Behaviour Therapy

Researcher: Samantha Grimes (Trainee Educational Psychologist) Email: Samantha.Grimes@birmingham.gov.uk

Telephone: 0121 303 8288. University supervisor: Neil Ryrie. neil.ryrie@nottingham.ac.uk

As part of my research into the social competence training that your child has

recently taken part in, I would like to ask them some questions about their

experience of the group sessions.

This will take no more than 15-20 minutes and will include questions such as

‘Do you think the ART group has helped you?’

This interview will be recorded on a voice recorder so that the information can

be referred to at a later date. Whilst excerpts from these recordings may be

used in the final write-up, great care will be taken to ensure that these are

anonymised and that any data gathered is stored securely.

Your child will be made aware, in their own consent form, that they are free to

leave at any point and can choose not to answer any questions.

If you are happy for your child to take part in the questions, please could you fill

out the attached consent slip and return it to the school office no later than 20th

January 2015.

If you have any questions or concerns please do not hesitate to enquire at your

child’s school or using the contact details above.

Child’s name____________________

Year___________________________

Please cross out as necessary:

Have you read and understood the information sheet YES/NO

245

Have you had the opportunity to ask questions and discuss the study YES/NO

Have all the questions been answered satisfactorily YES/NO

Have you received enough information about the study YES/NO

Do you understand that your child is free to withdraw from the study at any time without having to give a reason and that you are also free to withdraw on their behalf YES/NO

I give permission for the session to be recorded on voice recorder YES/NO

Do you give permission for your child to take part in the interview session (subject to them giving consent also) YES/NO

‘The interview has been explained to me to my satisfaction, I give permission for

my child to participate. I understand that my child is free to withdraw at any

time.’

Signed: ____________________________________

Print name: _________________________________

Date: ______________________________________

246

7.18 Appendix XVIII: Pupil Information Sheet and Consent Form for

Participation in the Qualitative Measures

Pupil consent form for an Interview Session

Researcher: Samantha Grimes (Trainee Educational Psychologist), University

of Nottingham, School of Psychology.

As part of my research into the Aggression Replacement Training groups that

you have attended I would like to ask you some questions about what you

thought of the sessions, such as ‘Do you think the ART group has helped you?’

This should take no more than 15-20 minutes. For this last set of questions I will

use a voice recorder so that I can remember everything you said later when

writing up the results.

All of the information that you provide will be kept confidential (your name will

not be used), it will be stored safely and if you do not want to answer the

questions in the interview you can leave at any time.

Please cross out as necessary:

Have you read and understood the information above YES/NO

Have you had the opportunity to ask questions and discuss the interview YES/NO

Have all the questions been answered so that you understand what you are being asked to do YES/NO

Do you understand that you are free to leave the question sessions at any time without having to give a reason YES/NO

Do you give permission for Sam to record this session on a voice recorder? YES/NO

Do you agree to take part in the study YES/NO

‘The interview has been explained to me and I understand what I will be doing. I

agree to take part in Sam’s questions. I understand that I can choose not to

answer these questions at any time.’

247

Signed: ____________________________________

Print name: _________________________________

Date: _____________________________________

248

7.19 Appendix XIX: Tables displaying exploratory analyses of normal distribution

Table 7.3 Table to Show the Skew, Kurtosis and Shapiro Wilk Analyses for the Self Report Data from the Control Group

*scores in bold represent the which is not normally distributed.

pre-test

post-test

Dependent variable Skew SE z score Kutosis SE z score SW Skew SE z score Kutosis SE z score SW

Communication -0.499 0.580 -0.860 -1.000 1.121 -0.892 0.085 -0.074 0.580 -1.276 -0.275 1.121 -0.245 0.394

Cooperation 0.101 0.580 0.174 -0.955 1.121 -0.852 0.753 0.331 0.580 0.571 -1.413 1.121 -1.260 0.133

Assertion -0.044 0.580 -0.076 -1.321 1.121 -1.178 0.248 -0.097 0.580 -0.167 1.336 1.121 -1.192 0.331

Responsibility 0.180 0.580 0.310 0.309 1.121 0.276 0.986 -0.382 0.580 -0.659 0.101 1.121 0.090 0.673

Empathy -0.177 0.580 -0.305 -0.702 1.121 -0.626 0.938 -0.139 0.580 -0.240 -0.751 1.121 -0.670 0.610

Engagement 0.508 0.580 0.876 1.953 1.121 1.742 0.406 -0.511 0.580 -0.881 0.305 1.121 0.272 0.873

Self Control 0.276 0.580 0.476 -1.463 1.121 -1.305 0.112 0.587 0.580 1.012 -1.027 1.121 -0.916 0.101

Social skills raw -0.189 0.580 0.326 -0.442 1.121 -0.394 0.896 0.157 0.580 0.271 -0.188 1.121 -0.168 0.493

Social skills standardised -0.217 0.580 -0.374 -0.307 1.121 -0.274 0.829 0.137 0.580 0.236 -0.194 1.121 -0.173 0.700

Externalising 0.180 0.580 0.310 -0.542 1.121 -0.483 0.360 0.572 0.580 0.986 0.350 1.121 0.312 0.243

Bullying 0.255 0.580 0.440 -0.589 1.121 -0.525 0.938 0.777 0.580 1.340 0.127 1.121 0.113 0.338

Hyperactivity -0.071 0.580 -0.122 -1.268 1.121 -1.131 0.243 0.121 0.580 0.209 -0.638 1.121 -0.569 0.684

Internalising -0.110 0.580 -0.190 -1.160 1.121 -1.035 0.303 0.839 0.580 1.447 -0.107 1.121 -0.095 0.139

Problem Behaviours raw 0.251 0.580 0.433 -0.371 1.121 -0.331 0.280 0.315 0.580 0.543 -0.312 1.121 -0.278 0.865

Problem Behaviours Standardised 0.284 0.580 0.490 -0.369 1.121 -0.329 0.265 0.311 0.580 0.536 -0.329 1.121 -0.293 0.790

SRMS 1.158 0.845 1.370 -0.407 1.741 -0.238 0.055 -0.541 0.845 -0.640 -1.220 1.741 -0.701 0.564

249

Table 7.4: Table to Show the Skew, Kurtosis and Shapiro Wilk Analyses for the Self Report Data from the Intervention Group

pre-test

post-test

Dependent variable Skew SE z score Kutosis SE z score SW Skew SE z score Kutosis SE z score SW

Communication -0.893 0.752 -1.188 -0.414 1.481 -0.280 0.122 0.271 0.752 0.360 0.159 1.481 0.107 0.963

Cooperation 0.029 0.752 0.039 -1.074 1.481 -0.725 0.682 -0.034 0.752 -0.045 -0.927 1.481 -0.626 0.945

Assertion -1.241 0.752 -1.650 1.949 1.481 1.316 0.318 0.938 0.752 1.247 1.047 1.481 0.707 0.385

Responsibility -0.600 0.752 -0.798 1.765 1.481 1.192 0.583 -0.229 0.752 -0.305 -1.254 1.481 -0.847 0.714

Empathy -0.240 0.752 -0.319 0.761 1.481 0.514 0.205 0.089 0.752 0.118 -1.337 1.481 -0.903 0.608

Engagement -0.514 0.752 -0.684 -0.770 1.481 -0.520 0.381 -0.823 0.752 -1.094 -0.516 1.481 -0.348 0.400

Self Control 1.148 0.752 1.527 1.336 1.481 0.902 0.367 0.041 0.752 0.055 -0.239 1.481 -0.161 0.935

Social skills raw -0.802 0.752 -1.066 -0.826 1.481 0.558 0.108 0.012 0.752 0.016 -0.412 1.481 -0.278 0.989

Social skills standardised -0.686 0.752 -0.912 -0.997 1.481 -0.673 0.181 -0.020 0.752 -0.027 -1.263 1.481 -0.853 0.675

Externalising 0.295 0.752 0.392 -1.188 1.481 -0.802 0.594 -1.273 0.752 -1.693 -0.175 1.481 -0.118 0.008

Bullying 0.109 0.752 0.145 -0.803 1.481 -0.542 0.559 -0.572 0.752 -0.761 -1.563 1.481 -1.055 0.066

Hyperactivity 0.144 0.752 0.191 -0.329 1.481 -0.222 0.804 -0.745 0.752 -0.991 -0.600 1.481 -0.405 0.266

Internalising 0.564 0.752 0.750 -0.115 1.481 -0.078 0.725 0.807 0.752 1.073 0.020 1.481 0.014 0.519

Problem Behaviours raw -0.350 0.752 -0.465 -0.940 1.481 -0.635 0.564 -1.094 0.752 -1.455 -0.003 1.481 -0.002 0.038

Problem Behaviours Standardised -0.250 0.752 -0.332 -1.067 1.481 -0.720 0.646 -1.111 0.752 -1.477 0.032 1.481 0.022 0.042

SRMS 1.045 0.845 1.237 1.043 1.741 0.599 0.571 1.314 0.845 1.587 1.588 1.741 0.912 0.302

250

pre-test

post-test

Dependent variable Skew SE z score Kutosis SE z score SW Skew SE z score Kutosis SE z score SW

Communication -0.491 0.661 -0.743 -1.486 1.279 -1.162 0.029 0.385 0.661 0.582 -1.418 1.279 -1.109 0.225

Cooperation -0.132 0.661 -0.200 -1.821 1.279 -1.424 0.092 0.351 0.661 0.531 -1.918 1.279 -1.500 0.013

Assertion -0.019 0.661 -0.029 -0.366 1.279 -0.286 0.696 -0.011 0.661 -0.017 0.251 1.279 0.196 0.985

Responsibility 0.249 0.661 0.377 -1.634 1.279 -1.278 0.083 0.432 0.661 0.654 -1.441 1.279 -1.127 0.088

Empathy 0.409 0.661 0.619 -1.617 1.279 -1.264 0.041 0.239 0.661 0.362 -1.305 1.279 -1.023 0.431

Engagement -0.582 0.661 -0.880 -1.104 1.279 -0.863 0.156 -0.031 0.661 -0.047 -0.153 1.279 -0.120 0.749

Self control 0.154 0.661 0.233 -1.767 1.279 -1.382 0.086 0.922 0.661 1.395 -0.430 1.279 -0.336 0.069

Social skills raw -0.119 0.661 0.180 -1.955 1.279 -1.529 0.066 0.383 0.661 0.579 -1.322 1.279 -1.034 0.287

Social skills standard -0.166 0.661 -0.251 -1.980 1.279 -1.548 0.054 0.343 0.661 0.519 -1.429 1.279 -1.117 0.222

Externalising 0.200 0.661 0.303 -1.690 1.279 -1.321 0.064 -0.303 0.661 -0.458 -1.466 1.279 -1.646 0.268

Bullying 0.573 0.661 0.867 -1.539 1.279 -1.203 0.009 0.058 0.661 0.088 -1.626 1.279 -1.271 0.119

Hyperactivity -0.224 0.661 -0.339 -2.003 1.279 -1.566 0.028 -0.438 0.661 -0.663 -1.548 1.279 -1.210 0.074

Internalising 0.167 0.661 0.253 -1.737 1.279 -1.358 0.053 0.725 0.661 1.097 -0.448 1.279 -0.350 0.399

Autistic Spectrum -0.070 0.661 -0.106 -1.932 1.279 -1.511 0.089 -0.103 0.661 -0.156 -1.064 1.279 -0.832 0.604

Problem behaviours raw 0.033 0.661 0.050 -2.003 1.279 -1.566 0.038 -0.327 0.661 -0.495 -1.227 1.279 -0.998 0.253

Problem behaviours standard 0.034 0.661 0.051 -2.029 1.279 -1.586 0.027 -0.254 0.661 -0.384 -1.465 1.279 -1.145 0.206

Table 7.5: Table to Show the Skew, Kurtosis and Shapiro WIlk Analyses for the Teacher Report Data from the Control Group

251

Table 7.6: Table to Show the Skew, Kurtosis and Shaprio Wilk Analyses for the Teacher Report Data from the Intervention Group

pre-test

post-test

Dependent variable Skew SE z score Kutosis SE z score SW Skew SE z score Kutosis SE z score SW

Communication -0.226 0.752 -0.301 -0.956 1.481 -0.646 0.697 0.342 0.752 0.455 -1.533 1.481 -1.035 0.393

Cooperation 0.051 0.752 0.068 -1.964 1.481 -1.326 0.077 0.388 0.752 0.516 -1.204 1.481 -0.813 0.608

Assertion -1.033 0.752 -1.374 -0.313 1.481 -0.211 0.063 -0.133 0.752 -0.177 -1.731 1.481 -1.169 0.375

Responsibility -0.143 0.752 -0.190 -1.025 1.481 -0.692 0.627 0.191 0.752 0.254 -1.488 1.481 -1.005 0.397

Empathy -0.416 0.752 -0.553 -0.461 1.481 -0.311 0.456 0.689 0.752 0.916 -0.321 1.481 -0.217 0.078

Engagement -0.439 0.752 -0.584 -0.250 1.481 -0.169 0.952 0.412 0.752 0.548 0.383 1.481 0.259 0.680

Self control 0.090 0.752 0.120 -1.111 1.481 -0.750 0.451 0.339 0.752 0.451 -0.886 1.481 -0.598 0.758

Social skills raw 0.376 0.752 0.500 -0.378 1.481 -0.255 0.836 0.306 0.752 0.407 -1.501 1.481 -1.014 0.509

Social skills standard 0.302 0.752 0.402 -0.618 1.481 -0.417 0.824 0.168 0.752 0.223 -1.668 1.481 -1.126 0.537

Externalising -0.567 0.752 -0.754 -0.519 1.481 -0.350 0.247 -0.263 0.752 -0.035 -1.307 1.481 -0.883 0.222

Bullying 0.275 0.752 0.366 -1.374 1.481 -0.928 0.476 -0.386 0.752 -0.513 -0.448 1.481 -0.302 0.408

Hyperactivity -1.098 0.752 -1.460 0.902 1.481 0.609 0.208 -0.275 0.752 -0.366 -1.483 1.481 -1.001 0.343

Internalising -0.196 0.752 -0.261 0.047 1.481 0.032 0.941 -1.330 0.752 -1.769 1.668 1.481 1.126 0.076

Autistic Spectrum -0.204 0.752 -0.271 -0.789 1.481 -0.533 0.962 -0.615 0.752 -0.818 -1.481 1.481 -1.000 0.018

Problem behaviours raw -0.502 0.752 -0.668 -0.489 1.481 -0.330 0.655 -0.767 0.752 -1.020 -0.778 1.481 -0.525 0.319

Problem behaviours standard -0.142 0.752 -0.189 -1.509 1.481 -1.019 0.661 0.130 0.752 0.173 -0.975 1.481 -0.658 0.912

252

Table 7.7: Table to Show the Skew, Kurtosis and Shaprio Wilk Analyses for the Parent Report Data from the Control Group

Pre- test

Post- test

Dependent variable Skew SE z score Kutosis SE z score SW Skew SE z score Kutosis SE z score SW

Communication -0.676 0.661 -1.023 -0.005 1.279 -0.004 0.502 -0.838 0.661 -1.268 0.689 1.279 0.539 0.596

Cooperation 0.061 0.661 0.092 -0.472 1.279 -0.369 0.988 -0.531 0.661 -0.803 -0.429 1.279 -0.335 0.639

Assertion 0.15 0.661 0.227 -0.148 1.279 -0.116 0.506 -0.984 0.661 -1.489 1.392 1.279 1.088 0.473

Responsibility -0.128 0.661 -0.194 -1.105 1.279 -0.864 0.649 -1.647 0.661 -2.492 2.963 1.279 2.317 0.023

Empathy 0.085 0.661 0.129 -0.904 1.279 -0.707 0.677 -0.596 0.661 -0.902 -0.442 1.279 -0.346 0.3

Engagement -0.054 0.661 -0.082 -0.939 1.279 -0.734 0.627 -0.011 0.661 -0.017 -1.037 1.279 -0.811 0.798

Self control -0.26 0.661 -0.393 0.839 1.279 0.656 0.214 -1.558 0.661 -2.357 3.245 1.279 2.537 0.055

Social skills raw -0.473 0.661 -0.716 -0.781 1.279 -0.611 0.578 -0.257 0.661 -0.389 -0.283 1.279 -0.221 0.752

Social skills standard -0.427 0.661 -0.646 -0.71 1.279 -0.555 0.761 -0.287 0.661 -0.434 -0.199 1.279 -0.156 0.663

Externalising -0.143 0.661 -0.216 -0.905 1.279 -0.708 0.874 0.373 0.661 0.564 -1.164 1.279 -0.91 0.348

Bullying 1.02 0.661 1.543 0.344 1.279 0.269 0.12 0.906 0.661 1.371 -0.482 1.279 -0.377 0.071

Hyperactivity 0.062 0.661 0.094 -1.592 1.279 -1.245 0.27 0.677 0.661 1.024 -0.635 1.279 -0.496 0.235

Internalising 0.366 0.661 0.554 -1.097 1.279 -0.858 0.164 1.501 0.661 2.271 3.267 1.279 2.554 0.076

Autistic Spectrum 1.206 0.661 1.825 1.2 1.279 0.938 0.085 0.31 0.661 0.469 -0.207 1.279 -0.162 0.878

Problem behaviours raw 0.214 0.661 0.324 -0.999 1.279 -0.781 0.716 0.773 0.661 1.169 -0.07 1.279 -0.055 0.507

Problem behaviours standard 0.369 0.661 0.558 -0.808 1.279 -0.632 0.551 0.921 0.661 1.393 0.469 1.279 0.367 0.523

253

pre-test

Post-test

Dependent variable Skew SE z score Kutosis SE z score SW Skew SE z score Kutosis SE z score SW

Communication -1.774 0.845 -2.099 2.863 1.741 1.644 0.014 -0.831 0.845 0.983 0.208 1.741 0.119 0.341

Cooperation 0.828 0.845 0.98 -1.809 1.741 -1.039 0.024 -0.148 0.845 -0.175 -1.128 1.741 -0.648 0.801

Assertion -1.113 0.845 -1.317 0.533 1.741 0.306 0.248 0 0.845 0 -0.3 1.741 -0.172 0.96

Responsibility -0.666 0.845 -0.788 0.586 1.741 0.337 0.473 -1.697 0.845 -2.008 3.064 1.741 1.76 0.078

Empathy -1.589 0.845 -1.88 2.96 1.741 1.7 0.2 -0.707 0.845 -0.837 -0.648 1.741 -0.372 0.442

Engagement -0.15 0.845 -0.178 -1.357 1.741 -0.779 0.439 -0.663 0.845 -0.785 0.586 1.741 0.337 0.923

Self control 0.515 0.845 0.609 0.729 1.741 0.419 0.953 1.301 0.845 1.54 1.28 1.741 0.735 0.295

Social skills raw -1.041 0.845 -1.232 2.206 1.741 1.267 0.502 -0.444 0.845 0.525 1.649 1.741 0.947 0.718

Social skills standard -1.129 0.845 -1.336 2.213 1.741 1.271 0.525 -0.568 0.845 -0.672 1.751 1.741 1.006 0.6

Externalising 1.447 0.845 1.712 2.227 1.741 1.279 0.169 1.927 0.845 2.28 4.121 1.741 2.367 0.042

Bullying 2.149 0.845 2.543 4.92 1.741 2.826 0.01 0.811 0.845 0.96 0.158 1.741 0.091 0.468

Hyperactivity 0 0.845 0 -0.008 1.741 -0.005 0.998 0.668 0.845 0.791 0.455 1.741 0.261 0.77

Internalising 0.148 0.845 0.175 -2.484 1.741 -1.427 0.237 1.682 0.845 1.991 3.029 1.741 1.74 0.083

Autistic Spectrum 0.578 0.845 0.684 2.315 1.741 1.33 0.394 1.778 0.845 2.104 3.575 1.741 2.158 0.071

Problem behaviours raw 0.988 0.845 1.169 1.499 1.741 0.861 0.663 1.888 0.845 2.234 4.159 1.741 2.389 0.051

Problem behaviours standard 0.974 0.845 1.153 1.018 1.741 0.585 0.673 1.438 0.845 1.702 2.683 1.741 1.541 0.293

Problem behaviours standard 0.369 0.661 0.558 -0.808 1.279 -0.632 0.551 0.921 0.661 1.393 0.469 1.279 0.367 0.523

Table 7.8: Table to Show the Skew, Kutosis and Shapiro-Wilk Analyses for the Parent Report Data for the Intervention Group

254

7.20 Appendix XX: Tables displaying the Results of the Analyses of

Homogeneity of Variances.

pre-test

post-test

Dependent variable F df sig. F Df sig.

Communication 2.173 1, 21 0.155 0.002 1, 21 0.964

Cooperation 0.557 1, 21 0.464 0.16 1, 21 0.693

Assertion 1.396 1, 21 0.251 1.562 1, 21 0.225

Responsibility 0.107 1, 21 0.746 0.006 1, 21 0.939

Empathy 1.777 1, 21 0.197 0.046 1, 21 0.832

Engagement 1.345 1, 21 0.259 0.014 1, 21 0.906

Self Control 4.846 1, 21 0.039 1.382 1, 21 0.253

Social skills raw 0.372 1,21 0.548 0.067 1, 21 0.789

Social skills standardised 0.309 1, 21 0.584 0.201 1, 21 0.659

Externalising 0 1, 21 0.992 0 1, 21 0.996

Bullying 0.019 1, 21 0.893 0.6 1, 21 0.447

Hyperactivity 2.157 1, 21 0.157 0.011 1, 21 0.916

Internalising 0.425 1, 21 0.522 0.02 1, 21 0.889

Problem Behaviours raw 0.033 1, 21 0.857 0.026 1, 21 0.874

Problem Behaviours Standardised 0.033 1, 21 0.857 0.026 1, 21 0.874

SRMS 0.225 1, 10 0.645 1.376 1, 10 0.268

Table 7.9: Table to Show the Non-parametric Levene’s test Results for the Self Report

Measures

Scores in bold represent data which was significantly heterogeneous in

its variances

255

pre-test

post-test

Dependent variable F df Sig. F Df Sig.

Communication 1.246 1, 15 0.282 0.234 1, 15 0.636

Cooperation 0.017 1, 15 0.898 0.72 1, 15 0.409

Assertion 0.152 1, 15 0.703 0.852 1, 15 0.371

Responsibility 10.715 1, 15 0.005 3.802 1, 15 0.07

Empathy 0.33 1, 15 0.574 0.009 1, 15 0.926

Engagement 0.49 1, 15 0.495 0 1, 15 0.994

Self control 0.001 1, 15 0.978 0.214 1, 15 0.65

Social skills raw 0.054 1, 15 0.819 0.356 1, 15 0.56

Social skills standard 0.054 1, 15 0.819 0.314 1, 15 0.584

Externalising 0.21 1, 15 0.653 2.295 1, 15 0.151

Bullying 0.244 1, 15 0.629 0.386 1, 15 0.544

Hyperactivity 1.573 1, 15 0.229 1.312 1, 15 0.27

Internalising 0.33 1, 15 0.574 0.856 1, 15 0.37

Autistic Spectrum 0.061 1, 15 0.808 0.056 1, 15 0.816

Problem behaviours raw 0.04 1, 15 0.844 0.874 1, 15 0.365

Problem behaviours standard 0.04 1, 15 0.844 1.461 1, 15 0.246

Table 7.10:Table to Show the Non-parametric Levene’s test Results for the Parent Report

Measures

pre-test

post-test

Dependent variable F df sig. F Df sig.

Communication 2.455 1, 17 0.136 0.051 1, 17 0.824

Cooperation 2.65 1, 17 0.122 0.536 1, 17 0.474

Assertion 5.755 1, 17 0.028 0.468 1, 17 0.503

Responsibility 0.503 1, 17 0.488 0.364 1, 17 0.554

Empathy 1.283 1, 17 0.273 9.576 1, 17 0.007

Engagement 1.938 1, 17 0.182 0.159 1, 17 0.695

Self control 0.995 1, 17 0.333 0.078 1, 17 0.783

Social skills raw 3.075 1, 17 0.098 0.186 1, 17 0.672

Social skills standard 3.155 1, 17 0.094 0.185 1, 17 0.672

Externalising 0.851 1, 17 0.369 3.279 1, 17 0.088

Bullying 6.99 1, 17 0.017 7.108 1, 17 0.016

Hyperactivity 1.587 1, 17 0.225 3.349 1, 17 0.085

Internalising 1.588 1, 17 0.225 0.319 1, 17 0.579

Autistic Spectrum 0.314 1, 17 0.582 1.156 1, 17 0.297

Problem behaviours raw 1.279 1, 17 0.274 0.692 1, 17 0.417

Problem behaviours standard 2.72 1, 17 0.117 2.19 1, 17 0.157

Table 7.11: Table to Show the Non-parametric Levene’s Test Results for the Teacher Report

Measures

256

7.21 Appendix XXI: Image of the Original Thematic Network for the

Facilitator Questionnaire Data

Fig. 7.2 Original Thematic Network for the ‘Reported Outcomes’ Overarching Theme for the

Facilitator Questionnaire Data

257

Fig 7.3 Original Thematic Network for the ‘Factors Impacting Upon Success’ Overarching

Theme for the Facilitator Questionnaire Data

258

7.22 Appendix XXII: Image of the Original Thematic Network for the

Group Member Interview Data

Fig 7.4: Original Thematic Network for Both Overarching Themes Constructed from the Group

Member Interview Data

259

7.23 Appendix XXIII: Needs of the Role Model and Target Individuals

Provided by the Member of Staff from Each School Responsible for

Selection.

Target pupils Role models

Rude to staff

Refusal to follow instructions

Aggressive and defiant

Self esteem issues

Low level disruption

Externalised aggressive behaviour

Verbal conflicts with staff and peers

Physical fights

Verbal aggression to peers and staff

Physical violence with peers

Disruptive

Verbally abusive

Bullying

Theft

Defiant

Disrespectful/Rude to staff

Easily lead

Loses temper

Silly, easily wound up by peers

Physical aggression

Anger issues when younger

Lacks tact

Quiet and reserved

Disruptive

Emotionally immature

Fighting

Physically/verbally abusive

Uncooperative

Mischievous

Gets involved in other’s disputes

Table 7.12 Behaviour Descriptors for the Role Models and Target Pupils Provided by School

Staff During the Selection Process.

260

7.24 Appendix XXIV: Glossary of Abbreviations

Abbreviation Translation

ACT Anger Control Training

ART Aggression Replacement Training

BESD Behavioural, Emotional and Social Difficulties

CBT Cognitive Behaviour Therapy

EP Educational Psychologist

EPPI Evidence for Policy and Practice Information and Co-

ordinating Centre

EPS Educational Psychology Service

iCART International Centre for Aggression Replacement Training

RCT Randomised Control Trial

SCED Single Case Experimental Design

SD Standard Deviation

SEN Special Educational Needs

SENDA Special Educational Needs and Disabilities Act (2001)

SRMS Sociomoral Reflection Maturity Score

SRM-SF Sociomoral Reflection Measure-Short Form

SSIS-RS Social Skills Improvement System-Rating Scales

Table 7.13 Table of Abbreviations